Go to previous article.
Go to next article.
Return to 1998 Conference Table of Contents
Neil G. Scott
J. B. Galan
The Archimedes Project
Center for the Study of Language and Information
Stanford, CA 94305-4115
Computers are becoming ubiquitous and anyone who has difficulty accessing them may be severely disadvantaged. Not so long ago, a person could opt not to upgrade to a new program or device if it impacted their access and simply continue to use an older accessible version. Increasingly, users no longer have the option of using an older version. For example, documents produced on a new version of a popular word processor cannot be opened by an earlier version of the same program. To remain competitive in the modern workplace, all workers must be able to upgrade to the same version of any tools that are used throughout an organization. All too often, disabled workers fall by the wayside when they are unable to upgrade at the same time as their peers. The most striking example of this is the impact of the graphical user interface (GUI) on blind computer users. This group fared quite well while there was the option of using text-based programs. Since the GUI began to dominate the workplace, however, it has created almost impenetrable barriers for blind and visually impaired computer users.
The problem is not that the GUI is used in the workplace, it is that there is no satisfactory alternative to the GUI for individuals who can't see the screen. In this instance, the software has become too complex for traditional solutions in which special access is achieved by modifying certain parts of the program or by retrofitting it with simple add-on modules. This problem is compounded by the short active life of many software products. GUI-based operating systems are evolving so rapidly and are so complex that manufacturers of screen reading programs are barely able to bring a product to market before the operating system is overtaken by a newer version.
While several of the leading software companies, notably Microsoft and Sun, have adopted a very responsible attitude towards producing accessible products, there are still far too many companies that give only lip service to accessibility or else ignore it altogether. In spite of the domination of Microsoft and Sun in the workplace, there are millions of computers and many software products from other sources that have poor or non-existent access.
There are several different views about where the responsibility for access lies. Some groups believe that it lies with equipment and software manufacturers. Others believe that it lies with the user or rehabilitation organizations. For the following reasons, we believe it lies somewhere between these two extremes.
If it were possible to measure the accessibility of all existing products, we would find that every product falls into one of three categories: 1) products that are inherently accessible to all disabled people, 2) products that are accessible to some disabled people and can be made totally accessible through retrofitting, and 3) programs that are inaccessible to all disabled people and can't be made accessible through retrofitting.
In a perfect world, all products would fall into the first category. In the real world, however, there are very few products in the first category and far too many in the third. Some disability advocates believe that legislation should be used to force all manufacturers to create only category-one products. This is patently absurd because different disabilities generate many mutually exclusive needs. The real-world situation is that it is almost impossible to create category-one products but all products could be made to fit the second category. We believe there are two steps to achieving this. First, every product should incorporate all of the access features that can be sensibly included within practical constraints such as the intended market, operational complexity, size, cost, and so on. Second, every product should include a standard interface that can easily connect to external devices that enable individuals with particular disabilities to access all features of the product. Defining and creating this standard interface and the necessary external access devices is the main focus of our research within the Archimedes Project.
The Total Access System is being developed by the Archimedes Project as a way to provide easy-to-implement and easy-to-use alternatives for people who need to use products that are otherwise inaccessible. It consists of three basic components: 1) a standardized interface to the computer or electronic device that is to be accessed, 2) a personal access device that matches the needs of the user, and 3) a universal method for interconnecting the first two components. Let's look at each of these in turn.
A TAP is a small electronic device that connects to any computer or electronic device that is to be accessed and provides a standardized interface to all of the controls and displays available to a normal user. The TAP functions by translating between the particular electronic codes that carry information within the target system, and a standardized language that conveys information to and from a user's personal access device. Connecting a TAP to a target system has no impact on its speed and does not modify to normal operation of the system in any way. A different version of the TAP is required for each type of target system. Currently, we have developed TAPs for IBM PC, Macintosh, Sun, SGI and HP workstations. We also intend to develop TAPs for non-computer devices such as appliances, household devices such as automatic door and window controls, vending machines, information kiosks, and so on.
Present versions of the TAP are limited to input functions and connect to the keyboard and mouse ports of the target system. The keyboard and mouse of the target computer are plugged into the TAP and continue to function in the normal manner.
We have recently begun the development of a VideoTAP that will retrieve textual and graphical information from the display of a target system. After some initial processing in the TAP, the retrieved information is transferred to an accessor that converts it to various formats that are accessible to blind and visually impaired users.
An accessor is a personal access device that can be used to control any device that is equipped with a TAP. The accessor functions by translating between the interface required by a particular user and a standardized language that conveys information to and from a TAP. Operation of the accessor is totally independent from the target system.
An extremely important concept of the Total Access System is that any accessor must be able to interact with any TAP. This means that a disabled individual can use the same accessor to operate many different target systems. Conversely, any target system that has a TAP can be used by and disabled person who has an accessor.
Accessors are designed to match the specific needs, abilities, and personal preferences of each user. They can be as simple or as complex as required by the individual. A single preferred interface can be used for every target system a person uses. This provides the user with a consistent operating environment. This reduces the amount of training required while the user is learning to use the system. It can also improve overall performance because the same operations are used to perform similar operations in many different applications. An accessor can evolve with the user, independently of the devices that are being accessed.
Accessors provide the first truly long-term solution to a disabled individual's access needs. Since they are totally independent of the target systems, accessors are not automatically made obsolete each time the operating system is changed, a new applications is added, or even if the target system is totally replaced. This property of the Total Access System has significant implications for children passing through the school system. By the time disabled students graduate, it is not unusual for them to have learned, used for a year or two, and then discarded five or six different accessible computers or accessible interfaces. At each transition there are physical, mental, or financial costs to the student, the parents, the teacher, the therapists and the various funding agencies. A single, well-designed accessor could function for a disabled person all the way through the education process and into the workplace. The flip side of this is that should it be necessary to change an individual's accessor because of changing abilities, it is not necessary to make any changes to the target systems.
Using an accessor that is separate from the target system also provides performance advantages when computationally demanding input techniques such as speech recognition are being used. Speech recognition software is normally run in the same target system as the applications software. This can lead to competition for CPU resources with the result that both the speech recognition and the application run at less than full speed. The Total Access System allows both programs to always run at full speed.
The Total Access Link is a data communications system that connects any accessor to any TAP. The current version of the link uses an RS232 serial cable to connect a single accessor to a single TAP. For situations in which it is necessary to use a single accessor with multiple target systems, we have developed a voice-operated switch-box that automatically routes the output of the accessor to the appropriate TAP.
This connection strategy is being extended in the next version of the Total Access System to allow multiple accessors and TAPs to share a single connection. A small local area network will connect all of the accessors and TAPs in a dynamic, programmable, user-controlled environment. This is an extremely significant development in the field of accessibility because it will eliminate the need for hardware configuration and enable disabled users to mix-and-match accessors to suit their individual needs and preferences.
When the user decides to work on a particular target computer, the total Access Link will connect the user's accessor to the appropriate TAP. The TAP will then transmit information to the accessor that allows it to automatically configure itself to match the characteristics of the target computer. It will be possible for more than one accessor to be talking to a particular TAP at any time. Multiple accessors will also be able to communicate with each other over the link to coordinate the performance of complex tasks This will allow the user to select and deselect access tools for each task or subtask. For example, a user might edit text by moving the mouse cursor with a head pointer and clicking the buttons by voice on a PC but use a foot switch to click the mouse button when using a Sun workstation.
The behavior of an accessor can b quickly and easily modified at any time either automatically, or in response to a user command. This allows a user to choose a particular style of interaction for each application on a particular target system or when moving between different targets systems. It will be possible to use a single accessor to simultaneously work with several target systems. The link will provide the accessor with the data it requires to reconfigure itself each time it connects to a different target system.
The new Total Access Link will also support a wide variety of communications media. The most common are twisted pair, power-line networks, radio frequency (RF), infrared (IR), fiber optic and coaxial cable. An advantage of the new system is that transceivers for the different media options are available from industrial control system vendors.
Most of our initial research was focused on the development of TAPs for commonly used computer workstations. This was primarily an engineering problem since the TAP must meet precise performance requirements for specific workstations. TAPs have been developed for IBM PC, Macintosh, Sun, SGI, and HP workstations. We jokingly refer to the TAPs as stealth technology since they are designed to be completely invisible to the computer to which they are attached. The first versions of the TAP provide access to all of the input functions on the target computers.
We have recently begun development of a VideoTAP that retrieves information from a target computer by capturing the video output signals and applying a variety of pattern recognition and character recognition processes. The VideoTAP is primarily intended to provide access for blind and visually impaired users but it will also be valuable for other types of disability because it will allow the accessor to locate screen objects like icons and the mouse cursor.
Stanford is licensing the TAP designs to a company that specializes in accessible technology for manufacture and marketing. All of the current TAPs are designed to operate outside of the target computer. We hope that manufacturers will eventually include a TAP as a standard interface on all electronic devices.
Many different types of accessor will be needed to match the needs of individuals with different disabilities. The only mandatory requirement is that each accessor uses the standard communications language required for the Total Access Link. Accessors can be as large or small, as simple or complex, or as cheap or expensive as is necessary to accommodate the needs, capabilities and preferences of each user. To differentiate the different types of accessors, we use a naming convention that includes the primary access method. For example, we are currently working with speech recognition accessors, head tracking accessors, and eye tracking accessors. Since we are able to use any notebook computer as an accessor, all of the existing computer access technologies can be easily incorporated into accessors. For example, we have demonstrated scanning accessors and Morse accessors by loading the appropriate access software into a notebook along with the software that connects to the Total Access Link. Most of our experience to date is with speech recognition accessors combined with head tracking.
We have developed a macro language called TAP SCRIPT that allows an accessor to translate short user commands into long and/or complex operations on a target system. TAP SCRIPT commands also provide the framework for coordinating the execution of complex tasks that use multiple accessors
Various versions of the Total Access System have been used regularly by almost one hundred people over the past three years. We have gathered a great deal of anecdotal evidence that the system works as expected. We are currently conducting a pilot study to determine the effectiveness of the system as a means for ameliorating the effects of repetitive strain injuries and other Musculoskeletal Disorders (MSDs). This will be followed by a formal study involving Stanford University, Hewlett Packard, and Boeing. We hope to publish the results of the formal study within a year.
In this paper, we have described the Total Access System and shown how it has the potential to make any device accessible to any disabled user. We have shown how the Total Access System can ease the problems of making all products accessible by giving manufacturers and consumers the choice of including access within individual products, or providing external access devices that can be easily connected to any product. We are not promoting the idea that the Total Access System should be the only access option available, rather, it should complement the other options
Go to previous article.
Go to next article.
Return to 1998 Conference Table of Contents
Return to Table of Proceedings