A Case Study
Topics in HCI: The Evolution of Ideas & Products
March 14, 2005
Professor Terry Winograd & Professor Donald Norman
In this paper, I will evaluate the evolution of screen reader technology. A screen reader is software that translates the contents of the computer screen to either speech output, using a Text to Speech (TTS) engine, or to a Braille display. For the purposes of this evaluation, I will focus on screen readers that use a text-to-speech engine. Screen readers with TTS output read the contents of the screen aloud and make it possible for visually impaired users to read and write documents, use the Internet, and send email using only a keyboard.
The market for screen readers is not considered mainstream – the screen reader is an adaptive technology for users with disabilities. Throughout its history, the driving issues have changed. While speech synthesis technology was not initially aimed at filling a need for blind users, extensive research in this field contributed to the development of future screen reader technology.
In evaluating screen reader technology, it is important to look at the broader realm of accessibility. Many within the disabilities community like to say that accessible technology is a market issue. The common opinion is that without improvements in access to technology for disabled users, there is an important consumer market that is neglected. It is estimated that about 1.1 million people in the U.S. are blind and each year 50,000 more will become blind. (NFB, 2004.) However, out of the estimated 295 million people in the U.S. (U.S. Census Bureau) this represents a relatively small market. Hence, conventional market economics have played less of a role in the development and adoption of screen readers.
While the history of assistive technology and screen readers is quite convoluted, one important factor that has started to and will likely continue to impact the development and adoption of accessible technologies is activism and the enactment of federal regulations. In 1998 Congress amended the Federal Rehabilitation Act to require that Federal agencies’ electronic and information technology be accessible to people with disabilities. The law also prohibits federal agencies from developing, buying or using electronic and information technologies that are inaccessible to people with disabilities. This law has important implications for accessible technology in general and in turn may prove to be an economic driving force for the advancement and adoption of screen reader technology.
We can look at the history of screen reader technology in terms of four key timeframes. These include the era before the PC, the period of the DOS-style PC, the period of the Graphical User Interface (GUI), and the era when Web technology became prevalent. There was an evolution in screen reader technology throughout that last three of these phases in order to continue to provide access to content on the screen for blind and visually impaired users. We’ll start by looking briefly at the origins of synthesized speech.
Origins of Text-to-Speech (TTS)
While the initial goals of research and development of speech synthesis technology were not necessarily to provide what we now call screen reader technology, it is important to look at the origins of speech synthesis and TTS.
Screen reader technology can be traced back to its roots in the development of speech synthesizers. At the 1939 New York World’s Fair, the first device to be considered a speech synthesizer, the VODER (Voice Operating Demonstrator) was introduced. The VODER was a manually controlled electronic speaking instrument invented by Homer Dudley, while at Bell Laboratories. The VODER was based on its counterpart, the VOCODER, also invented by Dudley, which was built as an attempt to save early telephone circuit bandwith by using speech compression. The VODER was operated by using touch-sensitive keys and a foot-pedal which controlled the electronic generating components. While speech quality and intelligibility were far from good and the device was difficult to operate, the VODER proved that speech could be artificially produced.
Before the PC
Research and development in speech synthesis technology continued over the next few decades in both academia and industry, with the invention of the first formant synthesizer, PAT (Parametric Artificial Talker), by Walter Lawrence in 1953, and the first articulatory synthesizer in 1958 by George Rosen of M.I.T., called the DAVO (Dyamic Analog of the Vocal Tract). (Klatt, 1987). The first full English text-to-speech system was developed by Noriko Umeda in the Electrotechnical Laboratory in Japan in 1968. (Klatt, 1987.) In 1979, the MITalk text-to-speech system was developed at M.I.T., followed by the Klattalk system developed by Dennis Klatt. The technology in the MITalk and Klattalk systems formed the basis for many of the synthesis systems today.
While much of the research in academia was aimed at understanding speech synthesis further and not necessarily focused on a potential for blind users, certain efforts were aimed specifically toward the blind. These included the Talking Typewriter developed by IBM Research in the 1960’s, the Braille Printer in 1975 and the Kurzweil Reading Machine for the blind in 1976.
Ray Kurzweil, an inventor whose interests in pattern recognition encouraged him to also invent the first omni-font optical character recognition (OCR), developed this reading machine for the blind. As the story goes, Kurzweil was “looking for a problem” for the OCR solution he’d found. He was sitting next to a blind man on an airplane who explained his needs of gaining access to various types of printed materials. Kurzweil found that there were two problems in the overall need – a flatbed scanner would be needed to scan in the books/printed materials and a full text-to-speech program would need to be developed to read the text after it was translated with OCR. He developed both of these products. The Kurzweil reading system was far too expensive for the individual consumer, (with a price between $30,000-$50,000), but they were used in libraries, schools and service centers for the visually impaired. (Leventhal, 2004.)
With the adoption of the personal computer in the 1980’s, screen reading technology for the individual blind consumer eventually became available. In 1988 (seven years after the introduction of the IBM PC with the MS-DOS operating system), the first screen reader for DOS was released. This product provided screen reading access for DOS using an external speech synthesizer. In these days, the screen reader directly read the 80X25 display buffer to speak the screen. The ASCII codes stored were sent to a hardware device to produce synthesized speech.
The screen reader for the PC opened up doors for employment for the blind. According to the National Federation of the Blind of North America, “Blind people could be found operating DOS based systems to do such jobs as order taking, word processing, customer service, accounting and more. For the first time in history, blind people could conceive of a world of equal access to the written word.” (Gabias, 2005.)
With the DOS screen reader many blind users could take advantage of computer technology. With DOS, speech output systems had relied on character-based computer displays. In the late 1980’s and early 90’s the widespread adoption of the Graphical User Interface (GUI) presented new issues for blind users. While an amazing advancement for sighted users, the GUI, which relied on icons and graphical representations, presented blind users with new barriers to access. Previously the 256 ASCII characters were fairly easily translated into speech or Braille using the DOS screen reader. Now, the pixel-based systems were presenting an entirely new paradigm and these systems were totally inaccessible to the blind.
Thousands of blind people in non-technical jobs had come to rely upon the computer. Many blind users continued to rely on the DOS system and waited to make the transition to Windows. However, it was just a matter of time before the trends in the workplace would force the switch. “Everyone from pizza order takers to medical transcribers could lose their jobs if no solution was found to the GUI problem.” (Gabias, 2005.)
Access to Windows for blind computer users was slow in development, but like the eventual access for DOS, it eventually emerged. The first screen reader for a Windows operating system was released in 1992. The first version of Windows had been introduced in 1985. That 7-year delay caused many in the blind community to criticize Microsoft’s lag in focusing on accessibility for the blind. In 1992, Syntha-Voice Computers released the SlimWare Window Bridge for Windows 3.1. At this time IBM also introduced a GUI screen reader, the Screen ReaderTM/2.
The difficulty at this point was the interoperability of separate screen reader applications and the Windows operating system. In Windows, the user interface control objects were not exposed, so the screen reader developers typically had to write special case scenarios for each application on the machine. The problem with this is that with each new release of software, there is the potential that the screen reader application would no longer work.
Disability Rights Activism
It is at this juncture that we start to see the impact of activism on the part of the disabilities and blind community. Advocacy organizations like the National Council on Disability and state agencies that serve the blind, such as the Commissions for the Blind in Massachusetts and in Missouri, urged Microsoft to commit to accessibility efforts for Windows 95 or potentially face refusals by those states to purchase the operating system when released. (Schroeder, 2000.) While technology access for blind users was not mandated by law at this point, the potential threat of losing state contracts or sales to these two state governments, might have had some impact.
Microsoft responded to the growing concern in the disabilities community and created a working group on accessibility. In 1993, the new director of Microsoft’s accessibility efforts, Greg Lowney, addressed the National Federation of the Blind in Computer Science, stating that "Windows has probably done more than anything else to earn Microsoft the enmity of the blind community. Microsoft has been both hated and feared by many people because we were promoting a graphical operating system without making sure that it could be used by people who are blind, and the results have been disastrous for many people." (Schroeder, 2000.)
Pressure from the blind community and several state agencies helped to influence Microsoft’s efforts to make future releases of Windows more accessible. They released Microsoft Active Accessibility (MSAA) in 1997, which exposes the user interface control objects so that assistive technologies can access them.
With MSAA, screen reader manufacturers can develop on a common platform, making the task of writing a screen reader easier and allowing access to the applications program and operating system.
Assumptions: Microsoft vs. Apple
Up to this point we have not mentioned accessibility issues related to Apple Computer. But it is important to note a bit of their history. Years before Microsoft created an internal group focused on accessibility, Apple had started its Worldwide Disability Solutions Group (WDSG), in 1985. The group, founded by Alan Brightman, had been seen as the industry’s most innovative team in the field of accessibility. Albeit a small group with five members, the WDSG worked with adaptive technology developers to try to make their existing hardware and software accessible. For a time, Apple was considered amongst the assistive technology community to be the computer of choice. (Clark, 2004.)
However, market forces came into play as Apple was losing money (on the scale of a billion dollars the previous year) and Steve Jobs dismantled the group in 1998 (reportedly saving 1 million dollars annually). “Apple was doing things to ensure disabled access when it wasn’t even on anyone else’s radar. But now that they don’t have people dedicated to working on the topic, they won’t be at the top anymore.” said Gregg Vanderheiden, director of Trace Research & Development Center at the University of Wisconsin. (Tedeschi, 1998.) Clearly this was a market issue and Jobs likely saw that the financial costs outweighed the benefits to Apple at that time. Perhaps his assumption was that accessibility was not going to be a profitable market and hence, did not justify the need for a dedicated team during a time of overall financial hardship within the company.
While Apple did continue to provide some accessibility features in its operating systems even without this centralized team, including zoom features for the visually-impaired and limited text-to-speech applications built into the system, one large concern amongst screen reader developers was the lack of Operating System hooks within the first release of OS X. These hooks, or background functions that programs such as screen readers can access directly, are important to ease the development of third party screen reader technology.
On the other hand, we discussed how Microsoft, with MSAA, had exposed the user interface control objects and provided these “hooks”, so that assistive technologies can access them in Windows. MSAA allows applications to expose screen readers to the type, name, location, and current state of all objects and notifies the screen readers of any Windows event that leads to a user interface change. While MSAA is not the only way for an application to communicate with assistive technology, it allows AT and screen reader developers to support a broader variety of applications without custom programming for each one. Microsoft also offers a set of tools that allow developers to see the information that MSAA is exposing – the MSAA Software Development Kit, making it easier for screen reader developers to work with Windows. It is difficult to know if Microsoft assumed a large market potential by incorporating this feature, but it put them in a position of control in the accessible operating system marketplace.
Today, there are about a dozen full screen readers currently on the market and the two that have emerged as leaders work only on the Windows operating system. This is not surprising given the ease of assistive technology development with MSAA vs. the difficulties caused by the lack of hooks in OS X. These two screen readers are JAWS (Job Access with Speech) by Freedom Scientific and Window-Eyes by GW Micro Inc. The cost of each of these products is comparable, which likely contributes to their steady competition, with JAWS running about $895 and Window Eyes at $795. “The two have been running neck-and-neck for years, trying to keep up with the other’s improvements”. (Speir, 2000) While this price seems fairly high given the likelihood of unemployment or lower income amongst the blind community, assistive technologies are notoriously expensive (because the market is smaller and the costs of development are shared by fewer people). Consumer choice is likely less dependent on cost, but more so on whether the device truly works. JAWS and Window-Eyes are used by “probably over 90% of users of the Web who are blind”. (Thatcher, 2002).
In addition to the standard desktop screen readers there have been efforts to create alternatives to the standard screen reader that are dependent on operating system access. These devices act independently of the operating system. One such device is the Total Access System, developed by the Archimedes Project, formerly affiliated with Stanford University. This system provides a GUI Accessor, which instead of using internal software, provides a “Visual TAP that recovers raw screen information directly from the video signals the target computer sends to its screen.” The goal is to provide access that is independent of the operating system (and hence what Microsoft or Apple controls) and to keep screen access as affordable as possible. While I have not found detailed figures on the rate of adoption for the TAS, the Archimedes Project web site states lists about 30 companies/organizations that are currently using the TAS, including Allstate, Boeing, Bank of America, Hewlett-Packard, Oracle and Sony among others. (Archimedes Hawaii, 2005.)
(Archimedes Project, 2005)
External Factor: Government Regulation
In 1998, Congress amended the Federal Rehabilitation Act to require that Federal agencies’ electronic and information technology be accessible to people with disabilities. “Section 508 was enacted to eliminate barriers in information technology, to make available new opportunities for people with disabilities, and to encourage development of technologies that will help achieve these goals.” (Section508.gov). This law also prohibits federal agencies from developing, buying, maintaining, or using electronic and information technology that is inaccessible to people with disabilities.
This law has important implications in the adoption of assistive technologies and screen readers, in particular. This law required that federal agencies purchase technology that is accessible to individuals with disabilities – i.e., any technology company that wants to sell their software, hardware or services to the federal government must ensure that their products are accessible.
Although Section 508 doesn't require private companies to alter their products, this law provides an incentive for companies who want to do business with the U.S. government to build better accessibility features into their products. Luckily for Microsoft, listening to accessibility proponents in the late 90’s and developing MSAA, meant that not only are the two most popular screen readers only Windows-compatible, but the Windows operating system, itself, was considered accessible without needing modifications to follow this new law. Their current market monopoly in general was not going to be negatively affected by this law, but perhaps enhanced. It is important to note that JAWS and Window-Eyes, which only work with Windows, are the leading screen readers used by federal employees. (Speir, 2000)
Another impact of Section 508 on the market for accessible technologies and screen readers is in relation to the Web. The law requires that individuals with disabilities, who are members of the public seeking information or services from a Federal agency, have access to and use of information and data that is comparable to that provided to the public who are not individuals with disabilities. (Access Board, 2000.) Section 508 created enforceable standards - “it establishes a complaint procedure and reporting requirements, which further strengthen the law.” (Thatcher, 2002.) This aspect is key in how it effects the development of web sites – the fundamental medium for disseminating information. “As early as 1995, technical accessibility issues centered around access to content on the web and web-based technology. Accessibility barriers were being created by the shift from a text-based environment to a robust, multi-media environment.” (Waddell, 2002.)
This means that Federal Government web sites must be accessible. While non-profit organizations and others that receive federal funding are not explicitly required under section 508 to ensure that their information technology is accessible, many non-government agencies that receive funding have worked to ensure that their public electronic information is made accessible. In terms of defining what is “accessible”, the World Wide Web Consortium (W3C) has taken the lead. Across the world, there are now public policies and laws protecting the rights of people with disabilities to access content on the Web. The W3C, an international consortium made up of over 400 members, promotes the interoperability of the web, sets guidelines and establishes common protocols for the web. The W3C Web Accessibility Initiative (WAI) provides guidelines for designers and developers to create content that is accessible by screen readers and in compliance with Section 508. This is the stable international specification that most developers look to in terms of accessibility.
While Section 508 mandates accessibility for the domain of the federal government, some have looked to interpretations of the Americans with Disabilities Act to encourage accessibility for private sector web sites. The ADA says that any place of public accommodation must be accessible to people with disabilities. The law, enacted in 1990, lists 12 categories including hotels, restaurants, shopping centers etc. Since the law was enacted before the introduction and adoption of the web browser, it does not include the web as a “public accommodation”. However, some have sought to sue, using the ADA, in order to encourage the private sector to make their websites accessible for screen reader users.
In 1999, the National Federation of the Blind (NFB) filed a lawsuit against AOL. The AOL client software was not compatible with screen reader technology and the NFB sought to “declare its service a virtual ‘public accommodation’ as per the ADA.” (Cisneros, 2000.) The NFB eventually dropped the suit after AOL took steps to adopt a corporate policy on accessibility, hired a blind person to direct its accessibility efforts, and released versions of AOL client software that have shown improvement in compatibility with screen access software. Another case that brought attention to the issue of web accessibility was a suit filed against Southwest Airlines by a blind man and an accessibility advocacy group called Access Now. The Judge in this case ruled that the ADA does not apply to public “virtual” spaces, but the case did once again bring the issue to the attention of the media and private sector.
While the threat of legal action has put the concept of web accessibility on the radar of the private sector, the potential market for blind consumers is still small enough that many companies do not yet see the need for making their web sites accessible. Blind users make up a relatively small percentage of the population with purchasing power. With the Web, screen reader users are now not only dependent on the interoperability of their operating system and screen reading software, but on the code of the individual web sites. Screen readers need to convert what is a two-dimensional to a one-dimensional text string, which is referred to as “linearizing” the page – the screen reader reads only the text between HTML tags and reads some of the tag attributes including the text in “alt” tags and “title” tags, for example. When code is sloppy (current browsers for the sighted are very forgiving of non-standard HTML), screen readers can have a difficult time reading the page. If correctly followed, the standards set in the W3C Web Content Accessibility Guidelines ensure that one’s web site is compatible with screen readers. However, as the old saying goes, “you can lead a horse to water, but you can’t make him drink”. While many designers know that these guidelines exist, they aren’t often mindful of the needs of blind users and of how their site reads through a screen reader.
Some web authoring tool developers have taken it upon themselves to ensure that designers and developers are given the tools to follow standards when developing their web sites. Standards are also set forth by the W3C, specifically in terms of web authoring tools – the Authoring Tool Accessibility Guidelines (ATAG). The goal of these guidelines is two-fold: to assist developers in designing authoring tools that produce accessible web content and to assist developers in creating an accessible authoring interface. Companies like Macromedia and Adobe, among others, have worked to add features into their authoring tools that provide shortcuts for accessibility features and templates that are accessible. Macromedia includes accessibility features in Dreamweaver and Flash (along with other products) and Microsoft FrontPage now includes an accessibility validation tool.
According to Jakob Nielsen in 2002, “In most projects, accessibility has fairly low priority because project managers underestimate the number of people who are impacted by design problems. They think that they are just losing a handful of customers, whereas in fact they may be turning away millions of customers, especially among senior citizens, who constitute a big and rich group that's getting more and more active online." (Mayfield, 2005.) The Nielsen Norman group released a software product, called LIFT NN/g that allows developers to validate pages for compliance with W3C accessibility and Section 508 guidelines as they code. The tool worked with Macromedia Dreamweaver (4.0 or MX) on both Windows and Mac. (Mayfield, 2005.)
While many validation tools have been released including a popular (but not terribly accurate) tool called Bobby and companies that develop software and authoring tools have made monumental efforts to include features for accessibility into their products, the web developer has ultimate control over the code they write. Macromedia works with web content developers and Flash designers to encourage them to think about how to design their rich applications to be accessible for screen readers. The director of the accessibility group hosts trainings for designers and developers and has an accessibility blog for web designers to learn about accessibility. As an intern in the accessibility group at Macromedia, I have created training applications to help designers learn how to use screen readers and incorporate accessibility testing into their design processes. In many ways, it is up to the designers to seek out this information, but the more that the big software companies advocate for accessibility, the more we will see web sites that can be accessed by screen readers. Also, by creating authoring tools that themselves are accessible and encourage developers to use standards when coding, these companies are ensuring that they will currently maintain a stake in the federal government and education market.
While we have looked at the history of screen reader technology in terms of four major timeframes, the evolution and adoption rates of assistive technologies are not necessarily sequential. Changes and advancements in mainstream technologies are outpacing the advancements in assistive technologies and, being smaller companies with less human and financial resources, screen reader companies may find it difficult to keep up. While mainstream technology continues to afford the average user with more conveniences, if the large software companies are not mindful of how their products interact with assistive technologies, each release of new software could force the assistive technology to no longer work.
The situation is indeed tenuous – we saw that historically Apple went from a position of leading the field in providing access to the dependent screen reader, to one where screen reader developers found it more difficult to build for Macs than for Windows. With MSAA, many screen reader developers focused their efforts on building for Windows, rather than the Mac. While there are screen readers for Mac users, including one called OutSpoken, by Alva, the two current leading screen readers operate only on Windows. MSAA has provided a fairly stable development environment for screen reader producers. It is difficult to know how changes in the release of Longhorn may affect the interoperability of current screen readers. With the release of Longhorn, MSAA will be replaced with a new API called UI Automation. Apparently MSAA will still be supported through an emulation layer but it may require a “wait and see” to be sure that performance of current screen readers are not impacted.
Riding the rollercoaster that is accessibility, Apple is also taking a turn to implement a full screen reader with the new OS X Tiger, to be released later this year. This screen reader is supposed to be a fully integrated, built-in screen reader, which would not require a third-party solution.
Unlike that of the mainstream consumer market, it is difficult to look at the adoption curve of screen readers in the same way. Cost, which is generally a crucial factor in the adoption of mainstream products, remains consistently high in the realm of assistive technologies. A smaller market and limited resources force AT users to carry more of the cost burden. Federal regulation has had some effect on encouraging the larger software companies to make their products accessible to assistive technologies, due to the magnitude of federal purchasing power. Some state that as people age, the market for screen readers and standardized products will grow. If this is the case, it will be interesting to see if solutions like the integrated screen reader (in OS X Tiger) and alternative solutions like the Total Access System (TAS) will increase in popularity. It is likely, however, that the rollercoaster ride will continue for the foreseeable future, so long as the market remains marginal.
Access Board, Section 508 of the Rehabilitation Act: Electronic and Information Technology Accessibility Standards, Technical Assistance to Ensure Successful Implementation, Retrieved from web on 3/10/05: http://www.access-board.gov/508.htm
Archimedes Hawaii. 2005. Visual TAS Retrieved from web on 3/14/05: http://archimedes.hawaii.edu/Visual_TAS.htm
Chong, Curtis. 2003. America Online: Is it Accessible Now? The Braille Monitor. Retrieved on 2/21/05. http://www.nfb.org/bm/bm03/bm0305/bm030509.htm
Cisneros, Oscar S., 2000. AOL Settles Accessibility Suit. Wired News. Retrieved from web site on 2/21/05.
Clark, Joe. Accessibility on the Mac: Trouble in Paradise. TidBits: All the News that’s Fit to Byte. Retrieved from web on 3/9/05: http://db.tidbits.com/getbits.acgi?tbart=06311
Electronic and Information Technology Accessibility Standards. Architectural and Transportation Barriers Compliance Board. Published in the Federal Register on December 21, 2000. 36 CFR Part 1194, [Docket No. 2000-01], RIN 3014-AA25. Retrieved from Access Board web site on 2/23/05.
Freedom Scientific, 2003. Company History. Retrieved from web site on 2/21/05 http://www.freedomscientific.com/fs_about/history.asp
Gabias, Mary Ellen. 2005. Opening Windows Closes Doors. National Federation for the Blind: Advocates for Equality. Retrieved from web on 3/14/05: http://www.nfbae.ca/publications/index.php?id=501
Horn, Paul, Senior Vice President IBM Research. Accessibility at IBM: An integrated approach. Retrieved from web site on 2/23/05
IBM Corporation. 2003. IBM's focus on accessibility. Retrieved from website on 2/21/05. http://www-306.ibm.com/able/product_accessibility/ibmfocus.html
IBM Corporation. 2003. History of Accessibility at IBM.. Retrieved from website on 2/22/05. http://www-306.ibm.com/able/access_ibm/history_past_01.html
Klatt, Dennis H., Review of Text-to-Speech Conversion for English. Journal of the Acoustical Society of America 82(3), September 1987, pp. v-793, 57 pp., Copyright 1987, Acoustical Society of America. Retrieved from the web on 3/12/05: http://www.mindspring.com/~ssshp/ssshp_cd/dk_737a.htm
Leventhal, Jay and Wehberg, Koert. The Great Screen Reader Race: A Review of the Two Leading Screen Readers, AccessWorld, September 2001, Vol. 2, No. 5. Retrieved 2/21/05 from: http://www.afb.org/afbpress/pub.asp?DocID=aw020502.
Leventhal, Jay. 2004. The Man and the Machine: An Interview with Ray Kurzweil. AccessWorld, Vol. 5, Number 5. Retrieved from the web on 3/14/05: http://www.afb.org/afbpress/pub.asp?DocID=aw050505
Maxey, H. David. Ed. 2001 Speech Synthesis Technology, Smithsonian Speech Synthesis History Project, Division of Information Technology and Society, National Museum of American History. Retrieved from web site on 2/21/05. (http://www.mindspring.com/~ssshp/ssshp_cd/ss_spsyn.htm)
Mayfield, Kendra. 2002. Reality Check for Web Design, Wired News. Retrieved from web site 3/13/2005. http://www.wired.com/news/technology/0,1282,55190,00.html
Microsoft, 2004. Gates Stresses Value of Digital Technology in Making Information More Accessible to People with Vision or Print Disabilities. Microsoft PressPass: Information for Journalists. Retrieved on 2/23/05. http://www.microsoft.com/presspass/features/2004/nov04/11-15LibrarySummit.asp
Microsoft, 2004. Narrator, Set Narrator Voice Options. Retrieved from web site on 2/25/05 http://www.microsoft.com/windowsxp/using/accessibility/narratorvoice.mspx)
National Federation of the Blind, 2004. About the NFB. Retrieved from web site on 2/21/05. http://www.nfb.org/aboutnfb.htm
Regan, Bob. Senior Product Manager for Accessibility, Macromedia. What is Accessibility? Retrieved from web site on 2/21/05. http://www.macromedia.com/macromedia/accessibility/gettingstarted/accessibility.html
Schroeder, Paul. 2000. A Brief History of Microsoft and Accessibility. AccessWorld. Retrieved from web on 3/14/05: http://www.afb.org/afbpress/pub.asp?DocID=AW010402
Section508.Gov, 508 Law. Retrieved from web site on 2/24/05. http://www.section508.gov/index.cfm?FuseAction=Content&ID=3
Speir, Michelle. August 7, 2000. Screen Readers Open Windows for the Blind. Federal Computer Week. Retrieved from FCW.com on 2/21/05: http://www.fcw.com/fcw/articles/2000/0807/cov-access3-08-07-00.asp
Thatcher, Jim. 2002. Web Accessibility for Section 508. Retrieved from web site on 2/23/05. http://www.jimthatcher.com/webcourse1.htm
Jim Thatcher, Paul Bohman, Michael Burks, Shawn Lawton Henry, Bob Regan, Sarah Swierenga, Mark D. Urban, Cynthia D. Waddell. Constructing Accessible Web Sites, Glasshaus. Birmingham, UK, 2002.
Tedesci, Bob. 1998. Apple Pulls Plug on Sick Kids’ Site. Wired News. Retrieved from web on 3/14/05: http://www.wired.com/news/print/0,1294,12351,00.html
“The Vocoder”, Homer Dudley's Speech Synthesisers, "The Vocoder" (1940) & "Voder"(1939). Retrieved from web on 3/14/05: http://www.obsolete.com/120_years/machines/vocoder/
U.S. Census Bureau. Retrieved from web site 3/13/05: http://www.census.gov/population/www/index.html
World Wide Web Consortium, Web Accessibility Initiative. Web Accessibility is a Marketplace Issue. Retrieved from web site 2/21/05. http://www.w3.org/Talks/WAI-Intro/slide7-0.html