BRAILLE DEVICES AND TECHNIQUES TO ALLOW MEDIA ACCESS MARCH 1992 Prepared by Daniel E. Hinton, Sr., Principal Investigator and Charles Connolly SCIENCE APPLICATIONS INTERNATIONAL CORPORATION 3701 N. Fairfax Drive, Suite 1001 Arlington, VA 22203 (703) 351-7755 1.0 SCENARIO Braille Devices and Techniques to Allow Media Access. 2.0 CATEGORY OF IMPAIRMENTS Persons with vision impairments. 3.0 TARGET AUDIENCE Consumers with Vision Impairments. Persons with vision impairments will benefit from enhanced access to media information services and computer systems. This scenario on advanced materials and technology for implementing Braille provides a means to disseminate information to consumers with vision impairments. In particular, it provides a better understanding of the technology available to produce Braille over the next three to five years. Policy makers, including national representatives, Government department heads, and special interest organizations. Policy makers can use this scenario to better understand the issues related to media access for persons with vision impairments. In addition it provides a point of departure for policy makers to understand how advanced technology legislative or regulatory funding priorities within Government programs can accelerate Braille output device development. Researchers and Developers. This group will benefit through a better understanding of the needs of persons with vision impairments and specifically their printed media communications needs. Understanding media access requirements will assist researchers and developers in designing Braille media access functions into their future products to meet the needs of persons with vision impairments. Manufacturers. Manufacturers will benefit through a better understanding of Braille device requirements, the potential market size and the existing Federal Government requirements for media access for persons with vision impairments which can be met by adding a Braille capability to their systems. 4.0 THE TECHNOLOGY Louis Braille published a dot system of Braille in 1829 based on a "cell" of six dots. He defined the alphabet, punctuation marks, numerals, and later a system for music using the 63 possible dot arrangements. Braille is read by running a finger over a character and sensing the raised dot pattern. Braille output devices in use today include a stylus on a pocket-sized metal or plastic slate (analogous to a pencil and clipboard), Braille writers like the Perkins Brailler (analogous to typewriters), and computer Braille devices. Printed Braille can be stamped on both sides of a page in a process called interpointing. This process saves paper and reduces the size of Braille books. The nominal specifications for Braille dot, Braille cell and Braille page dimensions are set by the National Library Service for the Blind and Physically Handicapped (NLS). The NLS certifies all Braille transcribers sponsored by the Library of Congress based on these specifications: Braille dots: -- Height for paper Braille, 0.019 inches, uniform within tran- scription; -- Base diameter, 0.057 inches; Braille cell: -- Center-to-center distance between dots, 0.092 inches; Corresponding dots of adjacent Braille cells: -- Horizontal separation, 0.245 inches; -- Vertical (down page) separation, 0.400 inches; Braille page: -- Standard size, 11.5 inches wide by 11 inches high. -- Minimum margin for binding side, 1 inch; -- Minimum margin for other sides, 0.5 inch; -- Minimum weight of paper, 80-pound; -- Paper must be thick enough so that, at worst, 10% of dots break the paper surface, but thin enough to permit uniform dots of proper height. The Perkins Brailler, made by the Howe Press of the Perkins School for the Blind, Watertown, Massachusetts, is a machine for embossing characters on paper. It is widely regarded as the standard for quality within the industry and has been used for over 100 years. It is capable of embossing 25 lines of 40 characters each, which is the page layout implied by the NLS standards. The term "Perkins" is used almost generically to refer to all Braille machines. It should be noted that even though 11x11.5 inch paper is standard, many rely on 8.5x11 inch paper because it works well with a slate and stylus. For paperless Braille, approximately 20 grams of force at 0.010 inches displacement, and 0.020-0.030 inches displacement without opposing force, may be a useful guideline for acceptable feel. Different technologies have different force-displacement characteristics which Braille-literate people must evaluate on a case-by-case basis. The Braille Authority of North America is the committee that sets standards for Braille code in the U.S., and all sanctioned Braille code is based on a 6-dot Braille cell: 2 columns of 3 dots each. Figure 1 shows the Braille alphabet. Nemeth Code, which is the standard Braille notation for mathematics, Computer Braille Code, and Textbook Braille 1 2 3 4 5 6 7 8 9 10 a  . . . . . b  .  . . . c   . . . . d   .  . . e  . .  . . f    . . . g     . . h  .   . . i .   . . . j .    . . k  . . .  . l  .  .  . m   . .  . n   .   . o  . .   . p    .  . q      . r  .    . s .   .  . t .     . u  . . .   v  .  .   w .    .  x   . .   y   .    z  . .    Figure 1. Grade 1 Braille Alphabet are all based on 6-dot Braille cells as is the literary Braille used for mainstream text translation. Some paperless Braille cells are produced in the U.S. with 8 dots per cell-- 2 columns of 4 dots each--but these are for compatibility in the European market. In Europe, the extra two dots are used to represent upper case letters and computer characters: control characters and extended ASCII characters. Eight- dot Braille cells could be adopted as the standard for computers in the U.S., but that seems unlikely for five reasons: 1. Six-dot Braille has a long and successful history in the U.S., and the cost of replacing Braille printers and paperless Braille displays in a short time would be prohibitive. 2. More lines of 6-dot Braille fit on a page than 8-dot Braille, and Braille already takes up several pages per printed page. An embossed Braille document takes about 15 to 20 times the volume of the same document printed, making it even less likely that 8-dot cells would be adopted for Braille books. This is based on the fact that it takes approximately three Braille pages per standard print page and a Braille page is 11x11.5 inches vs 8.5x11 inches for a standard text page. Also, a Braille page is 3 to 6 times thicker than a standard text page. 3. No single standard has emerged for special computer characters in 8-dot Braille. 4. One third more dots per cell would make an 8-dot Braille output device considerably more expensive than a 6-dot output device; the cost per dot generally dominates the total cost of Braille displays. 5. The "War of the Dots," which ended in 1918 with the choice of modified French Braille notation over American Braille and New York Point, has made Braille experts extremely cautious about making major changes in Braille notation. 5.0 STATEMENT OF THE PROBLEM Persons with visual impairments have limited real-time access to computer information because existing Braille output devices are expensive and can only display 20-80 characters at a time. In the U.S., voice synthesis devices are used by more visually impaired Americans than paperless Braille devices due to their lower cost. Paperless Braille displays are more common in Europe, where the Government generally pays for displays. Affordable paperless Braille is needed because voice synthesis does not allow the user to quickly review material as it appears on the monitor or printed page, including its format and structure. With the advent of large CD-ROMs with database libraries containing millions of print characters, and the increasing availability of information accessible by computer, persons with visual impairments need Braille displays that allow them equal access to the text displayed for sighted persons on the computer monitor. The best Braille displays now available limit persons with severe vision impairments to a single line of 20, 40 or 80 Braille characters. This makes it difficult to scan through text files and look for headings or jump from paragraph to paragraph. There is an urgent need for larger Braille displays to allow persons with vision impairments text access capability equivalent to that of sighted persons. Several factors influence the demand for Braille displays: the rate of Braille literacy is low among persons with vision impairments in the U.S., perhaps 20 percent. This is because, in part, visual impairments often set in with advancing age when it is more difficult to learn Braille. Most legally blind Americans are elderly. Also, age can adversely affect hearing, so there are older Braille-literate Americans who cannot use voice synthesis technology. The segment of the population with deaf-blindness with little or no residual hearing, regardless of age, also cannot benefit from voice synthesis technology. Some people cannot use Braille because they have reduced tactile sensitivity, as with diabetes, age, and occupations that callous hands. Overall, the largest demand for paperless Braille in conjunction with computers comes from people who can use voice synthesis technology but, because of the need to study, review and edit text, need to use paperless Braille. Many people with vision impairments want the capability to produce computer-driven Braille displays containing 3 or 4 horizontal lines of 80 Braille characters each. Others want 3 or 4 lines of 40-42 Braille characters. Many persons with vision impairments would be satisfied with a refreshable Braille display that simulates the 25-line Perkins Brailler page. However, size, weight, power, reliability and cost per unit will determine the maximum Braille page size. Researchers should focus their attention on identifying fresh approaches to producing the dots required to form the Braille characters within the space limitations imposed by the Braille specifications listed in Section 4.0. According to Noel Runyon, an engineer at Personal Data Systems and a Braille user, the critical factors that are easiest to overlook in the design of a full-page Braille display include: 1. Speed. Most reading is skimming, not sequential, cover-to-cover reading. Also, people can learn to read Braille as fast as sighted people read print. 2. Navigation. If display updates cannot occur in the blink of an eye, it is important to be able to "point" to a part of the display, evaluate it, and go to another page without waiting for the entire display to update, because everyone needs to flip through pages. Single characters must be individually addressable, and readers need a feel for where they are on a page. 3. Cursor location is critical on a computer display. 4. Application-specific devices are too restrictive to meet the broader communication needs of Braille users. For example, sequential output devices are awkward for most types of reading, whether their output is Braille or speech. 5. Humble things like dust can render laboratory successes almost useless in real homes and offices. 6. Graphics capability is a major justification for the use of a full-page display rather than a smaller display. 7. Battery power is highly desirable. 8. Noise is an important factor, especially in offices, libraries, and other public places. 9. Cost can make the difference between a device that is evolutionary and a device that is revolutionary. 10. Elderly and pre-employment-age people have generally received the least attention when developing new Braille technology, so they tended to be left out of the decisions that led to existing devices. 6.0 THE DEPARTMENT OF EDUCATION'S PRESENT COMMITMENT AND INVESTMENT The primary reason that Braille media access is a priority is because approximately 100,000 Americans with vision impairments use Braille for written communication. According to the 1988 National Health Interview Survey, 600,000 Americans between the ages of 18 and 69 have blindness or visual impairments severe enough to limit their employment opportunities, and that number rises sharply with age. This is an indication of the size of the population who could potentially benefit from Braille literacy. Although the number of visually impaired people under 18 is relatively small, they can learn Braille most easily and use it for the rest of their lives, thus they can gain the most from Braille literacy. The Department of Education, and its predecessor, the Department of Health Education and Welfare (HEW), have funded Braille device research and develop- ment over the past 20 years. With the advent of personal computers in 1975, HEW began to fund research and development of computer Braille output devices such as the TeleBrailler, and MicroBrailler. Currently, the development of Braille capability is a stated research priority of the Department of Education as follows:  The Electronic Industries Foundation (EIF) Rehabilitation Engineering Center's Technology Needs Assessment Paper, "An Inexpensive Refreshable Braille Display," points out a need for a "low-cost, reliable paperless Braille display mechanism." That report follows up on the recommendations of the National Workshop on Rehabilitation Technology, sponsored by EIF and the National Institute on Disabili- ties and Rehabilitation Research (NIDRR). The Workshop recom- mended making "information processing technology for access to print graphics, including computer access" the top technology priority for visual impairments.  Several of the funding criteria of the Department of Education's National Institute on Disability and Rehabilitation Research (NIDRR) are directed at the high unemployment and underemployment rate of persons with vision impairments and severely visually impaired populations. Most severely visually impaired Americans are unem- ployed. Larger and more affordable Braille displays would improve the educational outlook of blind individuals, promote Braille literacy, and improve employment opportunities and job retention among the Braille literate. Another stated priority, advanced training for the blind and visually impaired at the pre- and post-doctoral levels, and in research, would benefit greatly from improved Braille display technology.  The Panel of Experts for the Department of Education program sponsoring this study consists of experts from industry and Government, including members of the sensory-impaired community. Their consensus opinion was that developing a larger Braille display is the highest priority for persons with visual impairments.  One of the Department of Education's 1991 Small Business Innovative Research (SBIR) Program Research Topics is to develop or adapt communication devices for young children who are blind or deaf- blind. An affordable Braille display could be used for games that would help young children develop the skills needed to read and write Braille. A Braille display would also be of some use for tactile graphics, though an evenly spaced array of dots based on the same technology might be better.  The Department of Education's NIDRR Program Directory, FY89, lists the Smith-Kettlewell Rehabilitation Engineering Center, among many other tasks, as testing, developing, and/or evaluating a Braille display technology. 7.0 ACCESS TO COMMUNICATIONS MEDIA Many federal, state, and local laws which influence access for persons with visual impairments. The most important single law related to access for persons who are vision impaired is Public Law 101-336, enacted July 26, 1990. Better known as the Americans with Disabilities Act (ADA), this law has broad implications for all disabled Americans and establishes the objective of providing access to persons with disabilities to physical and electronic facilities and media. The other law that impacts technology for persons with visual impairments is Public Law 100-407-AUG. 19, 1988 titled "Technology-Related Assistance for Individuals with Disabilities Act of 1988." Also known as the Tech Act, this law established a comprehensive program to provide for technology access to persons with disabilities. The law defines assistive technology devices: "Assistive technology devices means any item, piece of equipment, or product system, whether acquired commercially off the shelf, modified, or customized, that is used to increase, maintain, or improve functional capabilities of individuals with disabilities." Braille technology clearly meets this definition for persons with vision impairments and should be exploited to increase the ability of persons with vision impairments to obtain access to printed media. Within the findings and purpose of this law, Braille technology can provide persons with vision impairments with opportunities to:  exert greater control over their own lives by making literacy possible;  participate in and contribute more fully to activities in their home, school, and work environments, and in their communities;  interact with nondisabled individuals; and  otherwise benefit from opportunities that are taken for granted by individuals who do not have disabilities. 8.0 POTENTIAL ACCESS IMPROVEMENTS WITH ADVANCED BRAILLE TECHNOLOGY Table 1 shows a sampling of the Braille technology currently available. The base price of adding paperless Braille to a computer is now about $5000. This high cost forces many persons with visual impairments in the U.S. to use voice synthesizers which costs about $1000. Braille embossers starting at approximately $1700 for the Braille Blazer, cost about three times that cost of text printers used by the sighted population. Advanced Braille technology offers persons with visual impairments the potential for dramatic improvements in access to books and periodicals stored in computer-readable form or scanned. For example, at least 800 titles are already available on CD-ROM and that number will probably increase rapidly in the years to come. Another important access improvement would be to computer-based telecommunications, including databases, electronic mail systems, computer bulletin board systems and mail order systems, all of which generally consider a computer screen as a single unit. One-line paperless Braille displays have been a cost compromise when compared to the speed and agility that a full-screen display could offer. It is often desirable to skim text for relevant information, whether that text is a computer's display, magazine or newspaper article, or book. When skimming, the field of the display needs to be as large as possible. The only practical alternative is Braille paper output, but relying on a Braille paper printer (priced for individual use) is slow and paper-intensive. A multiple-line paperless Braille display offers tremendous improvements in skimming speed and effectiveness over existing Braille printers and single-line displays. It also would have a great impact on the ability of persons with vision impairments to do research and academic study, which often requires reading and rereading information. 9.0 ADVANCED BRAILLE TECHNOLOGIES There are two major approaches to producing paperless Braille. The simplest approach is to apply constant power to keep each dot raised or lowered, but many of the technologies used to move dots require a substantial amount of power (50 to 100 milliwatts per cell). An analysis of the power available to a full page Braille display provides insight into the power that can be allocated to each Braille dot/cell. In older houses, standard electrical outlet can provide about 1200 watts of power. About 250 watts of that must be allocated to the computer controlling the display, leaving 950 watts for the Braille display. Assuming the display's power supply is 50% efficient, that leaves only 475 watts of power in the form the display can use. An 80-cell display with 6 dots per cell can allocate almost 1 watt per dot; 8 dots per cell lowers that to about 0.75 watts per dot. A standard Braille Table 1. A Sampling of Existing Braille Products Note: Prices range from 1989-1991 so they may not be comparable. Brand Name Manufac- turer Price System Description HARD COPY Perkins Brailler Perkins School for the Blind $395- $730 None Braille Writers, Manual and Electric Mountbatten HumanWare Inc. $2595- $3170 None Braille Writer, Electronic Index Braille Em- bossers HumanWare Inc. $2895- $16,900 IBM Braille Embosser Braillo 90 Braillo Nor- way AS $5795 IBM Braille Embosser Braillo 200 Braillo Nor- way AS $39,995 IBM Braille Embosser Braillo 400 S Braillo Nor- way AS $78,995 IBM Braille Embosser Romeo Brailler Enabling Technologies Company $2695- $3450 All Braille Embosser Marathon Brailler Enabling Technologies Company $11,500 All Braille Embosser TED-600 Text Embossing Device Enabling Technologies Company $37,500 All Braille Embosser Braille Bl- azer Blazie Engi- neering $1695 All Braille Embosser ATC/Resus 214 Printer American Thermoform Corporation $15,995 All Braille Embosser Versapoint- 40 Braille Embosser Telesensory Corporation $3795 All Braille Embosser/Translator Ohtsuki BT-5000 Braille/Prin t Printer American Thermoform Corporation $5180 IBM Apple Braille Embosser/Printer Duran Dots-40 Arts Computer Products Inc. $710- $1510 IBM Adapter to Convert Brother HR-40 Daisy Wheel Printer for Braille Printing Stereo Copy Developing Machine Matsumoto Kosan Com- pany $6250 None Braille Copier Thermoform Duplicators for Braille American Thermoform Corporation $1750- $2895 None Braille Copiers Plate Embossing Device PED-30 Enabling Technologies Company $62,500 None Braille Plate Embosser for Printing Houses TACTILE READING SYSTEM Optacon II Telesensory Corp. $3495- $3995 All Portable Tactile Reading System InTouch Telesensory Corp. $395 Mac Optacon II Accessory Soft- ware for Mouse Access Optacon PC Telesensory Corp. $395 IBM Optacon II Accessory Soft- ware/Hardware for Mouse Access ONE-LINE BRAILLE DISPLAYS Braille Display Processor Telesensory Corp. $3695 IBM Apple Paperless Braille 20 Cells Braille Display Processor BDP 21 Telesensory Corp. $3695 IBM Paperless Braille/Translator 20 Cells Braille Display Processor BDP 20 Telesensory Corp. $3695 Apple Paperless Braille/Translator 20 Cells Braille Interface Terminal Telesensory Corp. $3995 IBM Paperless Braille 20 Cells Navigator Telesensory Corp. $3,995- $14,995 IBM Paperless Braille 20,40,80 Cells VersaBraille II+ Telesensory Corp. $5995 IBM Portable Paperless Braille 20 Cells KeyBraille HumanWare Inc. $5025- $7025 Toshiba Paperless Braille 20,40 Cells Alva HumanWare Inc. $8,995- $14,495 IBM Paperless Braille 40,80 Cells Braillex IB80 Index Inc. $14,495 IBM Paperless Braille; 80 Cells New Ability Brailler Densttron Corp. $2995 IBM Paperless Braille (Soft Braille) 40 Cells BRAILLE NOTES/COMPUTERS Notex Index Inc. $5800- $7900 IBM Portable Braille Notetaking Device/Computer with 20- or 40-Cell Paperless Braille Personal Touch Blazie Engi- neering $5500 All Portable Braille Notetaking Device/Computer with 20- Cell Paperless Braille Braille 'n Speak Blazie Engi- neering $905 All Portable Braille Notetaking Device/Translator SpeakSys Blazie Engi- neering $149 IBM Braille 'n Speak Interface PocketBra- ille American Printing House for the Blind $905 All Portable Braille Notetaking Device/Word Processor Eureka A4 Robotron Access Products Inc. $2595 IBM Portable Talking Computer with Braille Keyboard Nomad Syntha Voice Com- puters Inc. $2295 All Portable Talking Computer with Braille Keyboard Option FOR THE DEAF-BLIND POPULATION AFB Tellatouch, MS 170 American Foundation for the Blind $595 None Typewriter Keyboard Controlling a Paperless Braille Cell for 1-Way 1-on-1 Communication DiaLogos Finnish Central Association of the Visually Handicapped None Braille Keyboard with Six Paperless Braille Cells Connected to a Typewriter Keyboard with 1-Line Dis- play (TDD) for 1-on-1 or ASCII or TDD Modem Commu- nication InfoTouch Enabling Technologies Company $4000- $4900 None Braille or Typewriter Key- board Connected to a Romeo Brailler and a Typewriter Keyboard with 1-Line Dis- play (Superprint TDD) for 1-on-1 or ASCII or TDD Modem Communication TeleBraille Telesensory Corp. $5500 None Braille Keyboard and 20-Cell Paperless Braille Display Connected to a Typewriter Keyboard with 1-Line Dis- play (Superphone TDD) for 1-on-1 or ASCII or TDD Modem Communication page with 6 dots per cell could allocate just under 0.08 watts per dot; 8 dots per cell lowers that to 0.06 watts per dot. An 80-cell by 25-line Braille display, which could provide full text access to an IBM-compatible personal computer screen, could allocate just under 0.04 and 0.03 watts per dot, for 6- and 8-dot cells, respectively. Without any blank lines, a page of Braille text could be expected to have an average of 2 dots raised per cell, so, if only raised dots require power, that would mean the typical power available per dot would be about 3 times the minimum values for a 6-dot cell (4 times the minimum values given for an 8-dot cell). Unless the display is being used for graphics, it would be unrealistic to expect all dots to be raised at once. On the other hand, it would be unwise to design a display so that raising all the dots would blow a fuse or trip a circuit breaker in the user's home or office. A compromise may be necessary for very large displays but it is desirable to have the capability to raise or lower all dots simultaneously. Applying continuous power to the actuators is impractical for many Braille display actuator technologies because even a one-line display would require more power than a wall socket can provide; far more than a portable battery system could tolerate. Therefore, many paperless Braille displays raise or lower dots and then lock them into position until another page is displayed. Historically, the locking and unlocking mechanisms have required little or no power except while displaying a new page. In practical operation, these locking mechanisms reduce average power consumption by several orders of magnitude. The problem with locking mechanisms has been that they increase mechanical complexity, which tightens the manufacturing tolerances. Reliable actuators for Braille cells are available today but most of them require a locking mechanism to avoid excessive power requirements. Even if power constraints could be ignored, some actuators' locking mechanisms double as a way of ensuring that dots are raised to a uniform height, which is a requirement for Braille. Designs that employ locking mechanisms update the display dot by dot, cell by cell or in small groups of cells. This minimizes the peak power consumption by decreasing the display update rate. Alternatively, a storage device could slowly accumulate energy from the power source and release it all at once; which is how a portable camera flash works. With the more energy-intensive actuator technologies, a tradeoff is necessary between display size and refresh rate of the display. The inherent size and weight of most Braille display technologies usually justifies slowing the display update time moderately. Portable Braille displays are almost certain to require tradeoffs in power vs refresh rate because both average and peak power capabilities of batteries are strictly limited by acceptable battery size, weight, and frequency of replacement or recharge. According to a 1990 Smith-Kettlewell study, a one-line Braille display that slides up and down a "page" provides some of the advantages of a full-page display. The study, by TiNi Alloys, Oakland, California, and Smith-Kettlewell, San Francisco, suggests that a six inch long virtual page can even create the illusion of a full size page of Braille. This work may lead to an alternative approach to providing the feel of a full page Braille device in a simpler and more reliable format. Solenoid electromagnetic actuator technology has been most often tried for producing Braille. Tight packing is needed for displays of useful size, even with coil assemblies and components fabricated with truly miniature solenoids. Historically, solenoids have been power-intensive (requiring locking mechanisms) and prone to failure with dirt from normal use (i.e., grease, skin cells, pollen, and even volcanic ash). This leads to reliability problems because cleaning 6000 solenoids regularly for a full page display would not be a realistic option. Covering the solenoids with a protective plastic membrane keeps the solenoids clean, but slightly moist fingers skip across plastic so a plastic surface is undesirable. Power requirements and interference between neighboring solenoids are also problems that must be overcome. Developments in superconducting materials, and in motor and solenoid miniaturization, may help to solve the problems associated with large electromagnetic Braille display fabrication. Metec (Stuttgart, Germany), EHG (Nordstetten, Germany), and Tiflotel (Calolziocorte, Italy) have each produced electromagnetic (solenoid) Braille cells that are scalable to multi-line or full-page displays. But the technology was less than successful because of a combination of reliability and repair problems. Power requirements may have also been a factor. Clarke and Smith International (Surrey, England), has produced small quantities of electromagnetic Braille cells, but they were limited to two-line displays. Novanik (Karlstad, Sweden) was working on a 42-cell 29-line electromagnetic display as of 1987, but its status is unknown. Generally, companies seem to have given up on using electromagnetic actuators for Braille. Smith-Kettlewell's proprietary design, described later in this document, is the exception. Piezoelectric benders, sold in the U.S. are used in all mass-produced refreshable Braille displays with more than a few characters. Called bimorphs, the benders can be made with any of several materials. Lead zirocnate and lead titanate ceramics seem to be the most popular for Braille cells but other piezoelectric materials include single crystals such as Rochelle salt and ceramics such as barium titanate. Piezoelectric materials flex in the presence of an electromotive force. The piezoelectric Braille cells made by Telesensory in Mountain View, CA, are considered by many visually impaired people to have the best feel of any Braille cell available in the U.S. The Tieman cell, another popular piezoelectric Braille cell, is probably made by Kogyosha (Tokyo) working with Braille Equipment Europe in the Netherlands. It is used in the Alva and Braillex displays, and possibly the KeyBraille and Notex displays. Metex, and possibly other electromagnetic Braille cell manufacturers, have switched to making piezoelectric Braille cells. The current state of the art in piezoelectric benders limits refreshable displays with horizontal benders to one or two lines, and only single-line displays are commercially available. The reason for the size limitation is that piezoelectric benders bend very little per unit length, so they have to be much more than an inch long to obtain the bending motion necessary to lift the Braille dots into place. Elinfa, in France, developed vertical piezoelectric benders for Braille cells, reducing the area required per dot by a factor of five. In theory, the Elinfa cell used in the Personal Touch, could be used to produce a very large display but, to date, producing Braille cells with horizontal or vertical piezoelectric benders has been expensive and labor-intensive. New manufacturing technologies may be needed to overcome this problem. Ceramic piezoelectrics tend to be brittle, and single-crystal piezoelectrics tend to have other undesirable properties. Single-crystal Rochell salt, for example, has the strongest piezoelec- tric effect known; but its dielectric properties have led to the use of ceramic piezoelectrics instead. The retail price of piezoelectric displays is $20-25 per dot, which would easily put the retail price of a full-page piezoelectric Braille display over $100,000 per unit. This is far outside the price range of the typical user. Five factors have enabled piezoelectric displays to dominate the paperless Braille display market. First, although driving piezoelectric cells requires on the order of two hundred volts direct current (VDC), the average current is low enough that they have very low net power requirements. The Telesensory Navigator's power supply would only allow a maximum of 0.023 watts per dot. Power consumption is so low that all dots can be raised continuously, thus eliminating the complexity and reliability problems associated with locking mechanisms. Piezoelectrics are actually their own locking mechanisms, requiring no power to stay in position except to cancel leakage currents. Also, low power consumption allows the option of portable battery power for small displays. Second, each dot has few moving parts, the bender and the dot shaft, in the case of telesensory's Braille cells and there is no friction-based locking mechanism. The result is that piezoelectric displays are relatively immune from dirt and wear, though dirt can cause dots to stick, requiring ultrasonic cleaning. Piezoelectric displays are reliable due to a minimum number of moving parts and minimal friction. Third, piezoelectric displays can provide fast display updates because they are energy-efficient. Fourth, piezoelectric displays are very quiet. They make just enough noise to let the user know an update has occurred. Finally, the dots can be closely packed and therefore come very close to the standard Braille dimensions. The next generation of full page Braille displays must be able to provide refreshable Braille for significantly less than $20-25 per dot. It is very unlikely that a full-page Braille display could be sold for much more than the $15,000 price of existing 80-cell displays. This means a 25-line, 40-character display with 6 dots per cell, would have to be produced at a cost of $2.50 per dot. A significantly lower price, perhaps $1 per dot, would open up a much larger market for full-page Braille displays and serve many more persons with vision impairments. Tactiles is working on a machine with very low cost self-locking dots (<$.10) and a travelling "printhead" similar to a dot matrix printer. The target price is under $4000. The reliability needed for every dot in a Braille display is significant. For example, if every dot in a full-line Braille display (480 dots) worked 99% of the time the display would be error-free once in 125 displays. If the dots were 99.99% reliable, the 80-cell display would be error-free only 95% of the time and the full-page display would be error-free about 55% of the time. At 99.9999% reliability, the one-line display would have an error every 2000 lines, but the full-page display would still have an error once in about 165 pages. This makes the design of a reliable full-page display difficult. Moving parts tend to make a device unreliable, but a Braille display must have 6000 independently moving parts, each with reliability much greater than 99.99%. Telesensory's Braille cells have been around for a long time and are extremely reliable, but there have been considerable differences in reliability among the manufacturers. Also, character- istics of the user such as sweaty hands and a tendency to eat potato chips, affect reliability. A major technology shift is required to design a full page Braille display to meet the media access needs of persons with vision impairments. This new technology shift would incorporate advanced materials and computer control technologies. Advanced materials and manufacturing technology may make it possible to implement several lines of Braille; perhaps a full 40-character by 25- line page of Braille output. An example of a technology improvement that could facilitate the implementation of full-page paperless Braille is large array controllers for liquid crystal displays (LCDs). These LCD controllers can control 64 high-voltage lines, on the order of 120-180 volts direct current (VDC) from a single chip and could be used to control 10 piezoelectric Braille cells. They may also be useful for switching electrorheological fluids, which will be described later. Other LCD controllers are available that could control 20 or more elements. Before discussing the newer technologies, a review of earlier attempts at a full-page paperless Braille display is useful to prevent repeating mistakes. Historical Developments Thermostat metals were used in Braille Inc.'s prototype Rose Braille Display Reader, a full-page display patented in 1981 but was never commercially produced. The Texas Instruments thermostat metals used are bimetallic strips that bend when heated. In the Rose Reader, the shaft of each Braille dot has a grooved ring around it. The shaft would be pushed up by a spring, but a hook on the end of a bimetallic strip catches the groove on the ring and restrains the dot. When heat bends the bimetallic strip away from the ring, the spring raises the dot. A separate manually activated mechanism pushes all of the dots down again and signals the machine to display the next page. The unit included a panel of 12 control buttons and a cassette drive for storing text. According to Leonard Rose, the one prototype that was built had some dots that did not work because the device was handmade; possibly machine-made parts would have been more reliable. Unfortunately the only accurate way to measure reliability very near 100% would be to manufacture parts. This would require a considerable investment. According to Mr. Rose, putting the system into production would cost about $750,000 with units eventually selling for as little as $7500. The thermostat metals used in the Rose Reader are less expensive than piezoelectric elements, and can be designed in modular units for easier repair. However, the number of moving parts per dot, the direct use of heat and friction, and the use of a manual mechanical reset mechanism are all potential sources of reliability problems. In principle, replacing the manual reset lever with an electric one is easy, but that is likely to affect reliability. The Rose Reader requires raising the temperature of the metal strips by 30 degrees Fahrenheit to overcome friction. This requires significant power, so the dots are raised one at a time, at a rate of 200 dots per second. If all dots are raised, the total time required for a full display is about 30 seconds, although the average Braille page would take approximately 10 seconds. Based on an estimate from the patent information, if a page could be displayed within one second, it would trip a 15 ampere circuit breaker, even with only one third of the dots raised. Each dot requires on the order of a watt-second of energy to be actuated. This is because the temperature changes required in thermostat metal actuators that have to overcome friction make them relatively energy- intensive. It is not clear whether sufficiently reliable mechanical locking mechanisms are available to circumvent higher energy requirements, but there seems to be a pattern. Energy-intensive actuators with lower materials costs tend to require locking mechanisms that cost as much to manufacture as more energy-efficient actuators for a given level of reliability. In the end, the prototypes with locking mechanisms are generally too costly to manufacture or too unreliable to sell. When the development funding for these devices runs out, the designs are shelved indefinitely. What is needed is an actuator that is energy-efficient enough to be used without a locking mechanism, yet costs less than $2.50 per dot or less and fits in a standard Braille cell profile. From the late 70's to the mid-80's, the American Foundation for the Blind (AFB) experimented with injection-molded arrays of 64 x 64 dots, manipulated one row at a time by a single row of 64 solenoids, one row at a time. Four of these prototypes were built with the combined capability of producing a full page of Braille or graphics. Graphics capability turns out to be a mixed blessing because evenly-spaced dots are incompatible with standard Braille dimensions but would be a major advantage of a full-page display over a one-line display. The system had three major advantages: the pins were mechanically latched into position so power consumption was moderately low (because the system was slow); the feel of the display was good; and since the system was modular, a mechanism for repair by replacement was provided. There were two major problems: the 64- step display update was slow enough to offer no great advantage over paper output, and the system was expensive. Ultimately, the cost of the mechanical system was its downfall, and the method was not recommended for further development. Shape memory alloys were the technology used by TiNi Alloys, to develop a 20-cell by 3-line prototype display of 8-dot cells. Shape memory alloys are nickel-titanium alloys that forcefully return to a preset shape when heated and are usually alloys of nickel and tin, or of copper, zinc and aluminum. In this case, TiNi used a nickel-titanium (nitinol) allow in the form of a wire, one inch long and 0.030 inches in diameter. When an electric current heats the wire, it shortens, pulling the shaft of a dot down against the force of a spring. Each dot has a small flexible piece of sheet metal with a hole big enough to let the shaft of the dot pass through it if the metal is lying flat, but small enough that it catches the dot if the piece of metal is angled. The metal's resting position is angled. When a dot is lowered, its shaft catches on the piece of sheet metal and pulls it down flat enough to let the shaft pass through. When the wire cools, it stops pulling the shaft of the dot down, and the dot tries to spring up again, but that pushes the sheet metal back to its angled position, catching the dot's shaft so it cannot move. When a plate pushes all of the pieces of sheet metal flat, the dots are then free to move and all raise, clearing the module. Stops were used to give the display adequate feel. The display was built of 4- cell units because modularity makes it easier to build and repair a display. Funded by the National Institute of Health, the project did not get past the prototype stage, though it received an Excellence in Design award from Design News, a respected journal for design engineers. The technology was capable of displaying a full Perkins Brailler page. In fact, the software was written for an 80-cell by 25-line display but there were reliability problems with the detent and release mechanisms, which required tight manufacturing tolerances. With further development, the technology might have become cost-competitive with other Braille cells, but the two-year grant ended in 1990. An important cost driver was making and attaching the special metal wire, although better techniques are now available. Power requirements were 50 watts instantaneous, or about 1 watt per dot. The dots were given that power, one module (32 dots) at a time, for 50 or 60 thousandths of a second, so the energy requirement per dot is 0.050 to 0.095 watt-seconds. That is much better than thermostat metals, but not as efficient as piezoelectric technology. Shape metal alloys were a valuable experiment for paperless Braille, but their cost and power performance seem unlikely to do much more than match piezoelectric technology. Current Developments For several years, Smith-Kettlewell has been working on a proprietary electromagnetic Braille cell technology funded by NIDRR. It is limited to displays of 80 characters or less, but the cost is estimated to be $20 per cell, which is below the cost of piezoelectric displays. Few details are available, but the technology has a fast refresh rate and has the potential to be used in portable systems operated from battery power. Smith-Kettlewell has passed the design on to a developer, though no estimated date for production was given. During the past six months to a year, Blazie Engineering, Street, Maryland, has been working on a pneumatic display that uses puffs of air to move tiny bearings supporting Braille dots. No product is anticipated until 1993. The device requires a spacing approximately 0.015 inches greater than the standard distance between Braille cells. The display is supposedly scalable to a full page. Preliminary cost estimates are as low as $5 per cell, which would be less than $1 per dot. The feel of the display is said to be solid, and it is expected that the present refresh rate can be increased. Power requirements are predicted to be low, and a 20-cell prototype has been built. The display has two moving parts per dot and can be cleaned by immersing it in liquid. Performance and cost predictions based on the prototype must be considered preliminary. Recent advances in sequential soft-copy Braille displays have been made by Tactilics, Inc. and Densitron Corporation. Sequential soft-copy Braille displays are essentially belts that move across a "window" while Braille dots are raised on their surface. Densitron has been selling prototypes of a 40-cell deformable plastic disposable belt device for $2995. Its lack of navigability would appear to limit its use. Tactilics' belt is made of hard, molded nylon cell sections which they indicate is long lasting and self-cleaning. They claim its bi-directional control makes it highly navigable and that it is a true realtime "monitor." Also, when battery powered, the unit may be used as a portable "book" and that its mixture of high and low tech is a price breakthrough. Two units will be introduced in mid-1992: 1 50-cell for $1200 and an 85-cell model for $1500. Future Developments What lies beyond the existing systems is impossible to predict with certainty because, though completely new technologies are seldom discovered, old ones are constantly revitalized by new computer capabilities, materials, and manufacturing processes. Sometimes older technologies suddenly become practical due to material or other technology breakthroughs. Some companies are unwilling to discuss technologies they are considering for paperless Braille. Blazie Engineering suggested three: magnetostriction, electrorheological (ER) fluids, and polymer gels. Magnetostriction is the property of some alloys that cause them to forcefully expand in a strong magnetic field. "Giant" magnetostriction, an expansion on the order of 0.15%, occurs in alloys of certain rare-earth elements. An alloy of iron with terbium and dysprosium is used in an actuator sold by Edge Technologies in Ames, Iowa. There are serious problems with using that technology for Braille. Rare-earth elements are not really rare, but they are expensive to purify. The alloys used must be in the form of a single crystal which is presently expensive to refine and produce. Finally, the effect of magnetostriction is too small to be used directly without long pieces of the alloy, which may be both voluminous and cost-prohibitive. Levers are being tried for converting some of the force from the actuators to linear displacements that would be adequate for Braille. A hard limit seems to be that the cost of an individual actuator is still not competitive with piezoelectric technology, and the cost of coils to produce the magnetic field, exceeds the cost of the special alloys as the actuators gets smaller. As with piezoelectrics, benders can theoretically replace levers as a way of trading force for increased movement, but the single crystals are brittle. No one knows if their tensile strength is such that the elements will break if used in benders. Power requirements are estimated at a maximum of 10 watts per dot, which is high, but locking mechanisms may allow power management strategies with low average power requirements. Other magnetostrictive materials exist, but it is not clear that any provide enough of an effect to be useful for Braille cells. Magnetostrictive ribbons, which are used in sensors, provide extremely high efficiency, but they lose most of that efficiency in strong magnetic fields, thus limiting their maximum expansion. Magnetostriction has its highest efficiency when the actuator is moving back and forth rapidly, though that problem might be solvable by making the dots move back down slowly, thus extracting a displacement as close to the maximum as possible. In summary, magnetostriction does not appear to be a cost- competitive technology for Braille cells, though this may change with future breakthroughs in materials technology. Fundamental research on electrorheological (ER) fluids is being conducted at the University of Michigan (UM), Ann Arbor. In 1988, a UM scientist made an important breakthrough in electrorheological fluids development. ER fluids thicken when a strong electric field is applied to them, on the order of 2000 volts per millimeter. Their consistency changes from liquid to something "more like Velveeta cheese." This allows hydraulic actuators to be constructed. ER fluids stop flowing while in a strong electric field, so, as a hydraulic fluid, they can selectively apply pressure to actuators. Per hydraulic switch, power requirements are lower than piezoelectric technology but a pump is required to supply the pressure for the hydraulics, and therefore their overall efficiency is unclear. The breakthrough at UM was to find an inexpensive ER fluid that does not contain water. Water content lowered the efficiency and predictability of previous ER fluids making them impractical for actuators. The fluids used at UM are inexpensive but the particles suspended in them tend to separate from the liquid. More expensive ER fluids do not have this problem. Three problems are likely to arise with ER fluid-based Braille cells: the use of a liquid, difficulty with modularizing a system with fluid lines, and fluid pump power and noise. Without modules, a large Braille display could be very difficult to build and repair. There may be ways to modularize a hydraulic display. Pump power and noise may not turn out to be an issue, but the use of a liquid seems likely to be a challenge to developers. The need for intense electric fields could be reduced by using narrow gaps. Overall, ER fluids may be feasible for Braille cell development in the immediate future. Polymer gels are another promising technology for full page Braille displays is polymer gels. Polymer gels collapse when exposed to intense light. These gels are being developed at the Massachusetts Institute of Technology's (MIT) Department of Physics and Center for Materials Science and Engineering. Under the proper conditions, gels can be induced to reversibly release a large portion of their liquid content. This is called collapsing, because releasable liquid content increases the volume of a gel by factors ranging up to 350 or more and multiplying their length, width and height by a factor of 7. In 1990, researchers at MIT induced a light-absorbing gel to collapse by heating it with a visible laser after having induced gels to collapse with exposure to ultraviolet rays, voltages on the order of 5 volts, and changes in the surrounding liquid's temperature, composition, pH, and salt content. The visible light has the advantage of safety and speed over ultraviolet radiation, as well as providing a controlled way to induce small temperature changes through a sealed container. The sealed container is necessary because, to reabsorb the liquid, a collapsed gel must be immersed in the liquid. The liquid and gel have to be separable to exploit the volume change of the gel, but gel reaction times below a second require gels significantly thinner than a human hair. To manage fine fibers, researchers in Japan have formed gels into sponges or bundles of fibers, but reaction times are still greater than one second. The leading light-sensitive gel researcher at MIT, Dr. Toyoichi Tanaka, estimates that strands of gel one thousandth of a millimeter in diameter, about the diameter of muscle fibers, would react as fast as muscle tissue. Doubling the diameter of a gel quadruples its reaction time, though; some gels react very slowly with modest increases in fiber diameter. So far, heating a gel with a laser is moderately power-intensive. Minimal research and development could conceivably reduce the power required by several orders of magnitude. The choice of gel material, the concentration and choice of light-absorbing material in the gel, and other factors could significantly reduce power requirements. According to Tanaka, red lasers work better than the violet-blue laser that was used to estimate power efficiency. Diode lasers with power efficiencies of over 30% are now available that emit red light. Though diode lasers with lenses cost on the order of $50 a set, and that is for lower power and in quantities of a thousand, if one or more lasers were scanned over a Braille dot array with a mirror, gel technology might make it possible to implement a reliable full-page Braille display with reasonable cost, size, weight, and power. Tight temperature control would be a potential problem (i.e., temperatures held within + one degree), but the MIT researchers have experimented with gels at room temperature without temperature regulation, and in water. The MIT group uses a low-cost gel, which is also encouraging. The temperature at which a gel collapses can be controlled by the proportions of two liquids into which the gel is immersed. Even if some temperature regulation turned out to be necessary, advances in solid-state Peltier effect heating/cooling might be applicable, balancing the power requirements of laser(s) with those of a temperature regulation system. It is too early to predict whether the feel of a gel-based Braille cell would be adequate, but the technology shows promise with additional research and development. According to "Tactile Displays for the Visually Disabled--A Market Study, July 1987," published by the Swedish Institute for the Handicapped, materials that expand with moderate heating have been tested for application to Braille displays, apparently without success. That reference does not indicate what material was tested, but it would not have been a polymer gel. Gels could be used for a phase change, but that phase change could not be accurately described as a transition from a solid to a fluid. The piezoelectric materials currently used in Braille cells are not the only ones available. A study of recent alternatives would be worthwhile. For instance, A.V.X., in Myrtle Beach, South Carolina, started selling lead zirconate titanate (PZT) in multiple layers around the beginning of 1991. It appears to be possible to get actuators made from this material for less than $20 each, in quantity, making multilayer PZT marginally competitive with existing piezoelectric displays. Layering reduces voltage requirements, but it is unclear whether the increased materials costs would be justified. A tough plastic film, called polyvinylidene difluoride (PVDF), sold by Atochemwith the trade name of Kynar, may also be useful for Braille cells. It can be used at very high voltages, compared to ceramic piezoelectrics, but its efficiency is lower than that of ceramics. It is apparently better suited to vibrating Braille dots than static ones, for reasons that include power efficiency, but the feel of vibrating Braille dots is not as good as the feel of static dots. Further study is needed to determine whether these and other materials are appropriate for use in Braille cells. In particular, their response at low frequencies must be taken into account. Superconducting magnets will eventually facilitate the miniaturization of solenoids because strong superconducting magnets can be fabricated in a small package without the overheating, even when densely packed. By definition, electricity running through superconducting materials produces no heat, which makes superconducting magnets incredibly energy-efficient. So far, the "high temperatures" required to use high-temperature superconductors are on the order of 300 degrees below zero Fahrenheit, but they are slightly higher than the temperature at which nitrogen, the principal component of air, liquefies. Liquid nitrogen, which is used to keep existing high-temperature superconductors cold, has been touted as being cheaper than beer, but anything that cold must be treated with extreme caution in devices developed for the general public. Liability issues could be enough to eliminate superconductive solenoids from consideration for Braille displays. Also, the known high-temperature supercon- ductors are brittle and somewhat expensive, so even the discovery of supercon- ductivity near room temperature, if that is possible, would not guarantee applicability to Braille cells. Superconducting materials applications are in the fundamental research stage and it will be 5 to 10 years before applications are marketed. An entirely different approach to paperless Braille uses electrodes to indicate the presence of a Braille dot with a tiny electric shock below the threshold of pain. This approach has potential for low cost, high speed, and small size, but experiments have not produced a design acceptable to the end users. Until technology is perfected, it cannot be considered a viable technology for Braille displays. Problems include reduced reading speed and very wide variations in skin resistance, both among different people and with sweat and other factors. An alternative approach to high reliability would be to use what are called "smart" materials. Smart materials combine sensors and actuators to react to special situations. In this case, smart materials might be able to provide high reliability with imperfect locking mechanisms by verifying that a dot has been raised or lowered. If not, an actuator could be triggered several times, allowing the reliability per trigger to be lower. Alternatively, the actuator could be vibrated to free any dirt that caused the failure. The smart materials approach is a compromise, attempting to avoid the high cost of piezoelectric technology which does not need a locking mechanism against the high cost of an extremely reliable locking mechanism. The smart materials approach would give a reliability boost to a mechanism that is already reliable. It would not be adequate with a mechanism that has a high failure rate. This is because if a dot fails to work after perhaps two or three tries, then the display's electronics would have to indicate an error. When a display operates properly, the error message should not appear except in the case of a catastrophic failure. The smart materials approach might also be used to adjust the overall height of the dots on the display. The cost of this feature may be prohibitive. At this time, shape memory alloys would probably be the best choice for testing along with low-cost piezoelectric-film sensors. Telesensory's Optacon II bears mentioning because the piezoelectric device provides access to printed text and the technology might be adapted to Braille in some way in the future. Based on ceramic piezoelectric technology, it uses 100 vibrating rods (5 columns by 20 rows) to present the image of letters from a small camera. Many persons with visual impairments find the Optacon II useful for reading print and interpreting graphics. It provides instantaneous results and offers great flexibility and portability. Like Braille, learning to use the Optacon II takes many hours of training and not everyone can become proficient in its use. The Optacon II is intended to supplement Braille, not replace it. Embossed Braille requires no special reading equipment or equipment costs, and it was also found to be easier and faster to read than raised letters. The Optacon produces raised letters when used to read print. It is important to note that the skin's sensitivity to fixed dots differs from its sensitivity to vibrating dots. Vibrating dots are not used for Braille because vibration temporarily reduces the skin's tactile sensitivity and actually reduces the ability to read tactile information. This is unfortunate because many piezoelectrics move at least an order of magnitude more efficiently near their resonant frequency. Vibrating dots also generate a buzzing sound, but that is a solvable secondary issue compared to the human factors problem. A detailed discussion of tactile graphics displays is beyond the scope of this document, but they are closely related to Braille displays. The big difference is that they generally use an array of evenly-spaced dots instead of the standard Braille spacing. These displays offer the advantage of graphics capability, but, in general, they cannot produce Braille with standard spacing. This document has concentrated on paperless Braille technology. However embossed Braille technology is also important. The primary technology for embossing Braille is solenoids, and this seems likely to continue for many years. The only commercial alternative has been the use of molten plastic containing magnetic material. The plastic is magnetically guided onto paper to form Braille dots. Until 1990, Howtek sold an embosser, the Pixelmaster, based on this principle, but it could only raise dots eight thousandths of an inch. Even twelve thousandths of an inch is barely tolerable for Braille, and far below standard dot height. So the Pixelmaster, which could also produce printed text, was considered appropriate only for tactile graphics and visible print. Technologically, the Pixelmaster could have probably produced Braille of normal height, but it was originally designed for advertising. The plastic "ink" was originally intended for producing more brilliant colors displays and not to produce Braille. Similar technology was tested in Japan and found to produce the proper height of dots, although dot shape was a problem. Plastic dots on paper are much more durable than conventional embossed paper dots, but it is unclear whether their feel can be made acceptable to the users. Dots that are too smooth are harder to read. Smith-Kettlewell is in the very early stages of developing a thermal embossing technology, said to offer the possibility of fast and silent paper Braille. Multiple copies of Braille text can be made with heated plastic sheets that conform to the shape of the original Braille page. This process is called vacuum forming. In larger quantities, printing press techniques can be used to emboss Braille, but copying Braille is still expensive. Even in quantities of a few hundred, a relatively fast Braille embosser still has advantages over Braille copying technology, including the feel of paper vs plastic. Sheets containing encapsulated ammonia are being used to produce some tactile graphics, but the special plastic sheets required are too expensive to be used as a substitute for paper Braille. They are used with thermal copiers, but have the feel problems associated with plastic too. 10.0 COST CONSIDERATIONS OF ADVANCED BRAILLE TECHNOLOGY As explained in the preceding section, advanced technology Braille must cost less than 20 to 25 dollars per dot to be cost effective when compared to existing technology. A substantial price reduction would be needed to make multiple-line displays marketable, and a 40-cell by 25-line full-page display would have to cost less than $2.50 per dot to be in the price range of existing market forces. Piezoelectric displays, the dominant technology at this time, show little hope of dramatically dropping in price in the near future unless a much less expensive material for them is found or new manufacturing techniques formulated. Thermostat metals, though cheaper, are not energy-efficient, making them a poor choice for displays. Highly mechanical approaches tend to be expensive, unreliable, or both. Shape memory alloys tend to be somewhat expensive, though moderately energy-efficient. Smith-Kettlewell's electromagnetic technology seems to be able to offer a significant price breakthrough, but it is limited to one-line displays. Soft Braille offers an apparently inexpensive alternative to refreshable Braille technology, but its costs may be misleading. Refreshable Braille makes it feasible to move to different points in a document without too much confusion, but Soft Braille is best suited to cover-to-cover reading which is not the reading pattern except for some pleasure reading. Also a disposable belt model must be judged on the basis of the cost and inconvenience of replacing the belt. On the horizon, magnetostriction seems to offer little hope for a price breakthrough at this time. On the other hand, ER fluids seem to have the potential of producing a low cost Braille display. However, there are significant engineering hurdles to be overcome. Polymer gels offer future hope of low-cost displays, though they are still in the basic research stage. New developments in piezoelectric materials merit further study. Superconducting solenoids and silicon micromachines are probably beyond the ten-year scope of this study, and it is not clear whether electrode-based Braille will ever have the feel demanded by persons with vision impairments. Based on the outlook for affordable full-page Braille displays, two compromise approaches should be considered. First, smart materials are a way of increasing the reliability of existing mechanical locking mechanisms: Eliminating the need for locking mechanisms might be better in the long run, but that may not be technologically feasible without the expense of piezoelectric displays. Second, further investigation into the effectiveness of a sliding one- line display is justified by the lack of compelling evidence that a full-page Braille display is technologically feasible in the next three to five years. However, research and development efforts should be devised to push the technologies described above beyond the laboratory. 11.0 COST BENEFITS TO PERSONS WITH SENSORY IMPAIRMENTS WITH EARLY INCLUSION OF BRAILLE Braille displays are presently costly, and therefore many persons with vision impairments currently use synthesized speech instead. More affordable Braille displays would give persons with vision impairments more options between speech and Braille, and increase literacy among persons with vision impairments. Lower-cost Braille displays would allow larger displays to be purchased per dollar. Perhaps the biggest short-run cost benefit would be the improved earning potential that a better Braille display could give blind workers, especially in the computer programming and office environments. In the long run, an affordable full-page Braille display would contribute to Braille literacy, education of the blind, and access to computers, empowering persons with visual impairments. 12.0 PRESENT GOVERNMENT INVOLVEMENT IN ADVANCED BRAILLE TECHNOLOGY At present, no Government programs are known to be supporting Braille display development, although virtually all previous development have been Government-sponsored. 13.0 ADVANCED BRAILLE TECHNOLOGY TIMELINE Paperless Braille technology has settled on piezoelectric actuators because of their reliability and energy efficiency. To compete with piezoelectrics, any new technology must provide reliability and energy efficiency at a lower cost per cell. As computers become increasingly portable and dependent on battery power, only the most energy-efficient Braille displays can be used in these portable units. As the most energy-efficient proven technology for Braille displays, piezoelectrics should be reevaluated to determine if new manufacturing technologies can lower the cost of piezoelectric displays. Smith-Kettlewell's new electromagnetic technology, already in the process of becoming a commercial product, may provide significantly less expensive one line displays. Expected in 1993 is Blazie Engineering's pneumatic display which may make full-page displays affordable. Recent Soft Braille displays offer a compromise between low cost and limited performance. Their price is revolutionary, but their capabilities are extremely limited for most applications other than reading for pleasure. Magnetostriction is not likely to be a major factor in Braille displays, but clever designs or new materials breakthroughs could overcome the cost barrier. ER fluids, with the potential for low cost and high energy efficiency, should be ready to begin development in 1992. Polymer gels should be available in actuators between 1993 and 1994. If their energy-efficiency proves to be high enough to make them practical, they should be evaluated for Braille displays. However, they are still in the basic research phase in 1991. Superconducting solenoids and silicon micromachines are beyond the ten-year time scale being considered by this study, and electrode-based Braille seems unlikely to be practical at this time. A smart materials approach, with shape memory alloys and piezoelectric film sensors, merits some consideration, but it offers only a limited chance of success in competing with piezoelectric displays. That approach is moderate-risk, moderate-gain, and probably moderate-cost, since some of the development has already been done. Experimenting with a one-line display that slides up and down a page, that may be compressed vertically, is a higher-risk approach, but it offers potentially very high gains if it makes a full-page display unnecessary. Cost should be relatively low in this approach, making it especially attractive. 14.0 PROPOSED ROAD MAP FOR INCLUSION OF BRAILLE CAPABILITIES The U.S National Aeronautics and Space Administration (NASA) and the Department of Defense (DOD) fund actuators for specific systems. Neither is in the business of developing advanced actuators for Braille. However, in a cooperative funding teaming arrangement for research and development on small lightweight actuators there probably would be a great deal of interest. For example, both the DOD and NASA fund research and development efforts for devices for the handicapped through small business and university innovative research grant programs. The Department of Education, as a first step in Braille cell development, should explore the possibility of a cooperative effort with both NASA and the DOD in the area of small actuators for use in full-page Braille cells. This would lead to a research and development program for small low power consumption actuators for use in braille display devices, robotics, and space applications. A program would then be established to develop single-cell actuators that could be used in braille displays. NASA or the DOD could sponsor the initial basic research under small business or university educational grants for cells of one to six elements. This could be followed up with a program by the Department of Education under a grant for a Phase I program to develop a single-line Braille display of 80 characters. Finally, a grant would be awarded for an 80 column by 25 line display. Key to each phase would be three to four research organizations competing for the next phase of the program. For example, NASA might start with four to five organizations for a one year concept design study at approximately $125,000 per year, for small low power actuators based on advanced technology. Designs would be sought for both single actuator and multiple cell designs. NASA would then select the three organizations with the most feasible designs for a second phase to integrate the individual actuators into a prototype Braille cell display and fabrication effort at $150,000 each for one year. The basic research effort would then be followed up with a Department of Education effort to integrate one or two of the organizations' actuator designs into low-power, full-page Braille displays over a two to three year program. The Department of Education and NASA would jointly fund the efforts and cooperate in the exploitation of the technology. Finally, the Department of Education would fund a production study and transition the devices to full scale development. 15.0 POTENTIAL PROGRAM SCHEDULE Figure 2 presents a potential program schedule for the development of a braille display unit using advanced technologies. The Department of Education could act as the program administrator, with NASA and the DOD providing basic research and development expertise at critical times throughout the development effort. Within five to six years a full-page Braille display could be on the way to full scale production.