 |
Questions? Call 888-624-8373 |
|
PDF CHAPTERS
your price: $2.80
select
|
|
|
 |
|
|
|
The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Page 395
On Selected Population Groups
Extending Knowledge Access to
Underserved Citizens
Wallace Feurzeig
BBN Systems and Technologies
Issues And Goals
If we are serious about making available the rich human and
information resources of the national information infrastructure
(NII) to all citizens, we have to address the great, and growing,
knowledge gulf between the ''haves" and "have-nots" in American
society. This disparity poses a real threat to our political and
social lives. It is at variance with our history and our democratic
ideals. Many citizens who are economically underserved are also
"informationally disadvantaged" (National Telecommunications and
Information Administration, 1995). The knowledge empowerment made
possible by the new information technologies must become available
to all. We must work toward democratizing access to effective use
of networking technology by all Americans who are capable of
benefiting from its use. We need to provide the underserved not
only with access to these potentially empowering information
resources but also support in their use through education,
training, and acculturation.
|
|
|
|
|
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 395
Page 395
On Selected Population Groups
Extending Knowledge Access to
Underserved Citizens
Wallace Feurzeig
BBN Systems and Technologies
Issues And Goals
If we are serious about making available the rich human and
information resources of the national information infrastructure
(NII) to all citizens, we have to address the great, and growing,
knowledge gulf between the ''haves" and "have-nots" in American
society. This disparity poses a real threat to our political and
social lives. It is at variance with our history and our democratic
ideals. Many citizens who are economically underserved are also
"informationally disadvantaged" (National Telecommunications and
Information Administration, 1995). The knowledge empowerment made
possible by the new information technologies must become available
to all. We must work toward democratizing access to effective use
of networking technology by all Americans who are capable of
benefiting from its use. We need to provide the underserved not
only with access to these potentially empowering information
resources but also support in their use through education,
training, and acculturation.
OCR for page 396
Page 396
Researchers at BBN and elsewhere have done preliminary work
along these lines to better understand the issues. Though neither
systematic nor comprehensive, this work identifies some key
research questions and suggests specific directions for more
substantial action-oriented research efforts.
Internet Use By Low-Income
Families
A recent study conducted from December 1994 to January 1996
probed the barriers, benefits, and perceived worth of the Internet
to six low-income urban families in Florida, a group representative
of the traditionally underserved and informationally disadvantaged
population (Bier, 1996). The researchers asked what these families
would actually do on-line given unrestricted Internet access in
their homes. Each family was lent a home computer, high-speed
modem, and printer; was provided with dial-up point-to-point
Internet access; and was given training on the use of the mouse and
keyboard. The computers were equipped with an interface security
program, an integrated productivity package, several educational
games, a typing tutorial, and a set of Internet utilities. Families
were taught how to communicate with each other electronically and
how to locate and acquire resources from the Internet. Additional
training and technical support were available on demand for the
duration of the project. Through interviews, visits, and telephone
and e-mail interactions, researchers obtained data on the amount of
time participants spent on-line, the sites they visited, the
information they sought, and the obstacles they encountered. The
participants made use of virtual hospitals, medical dictionaries,
and physicians' desk references. They joined support groups,
investigated scholarships, and made local transportation
arrangements. They investigated appliances, employment listings,
and local calendars of events. They e-mailed, chatted, and surfed
the World Wide Web; made friends; felt personally empowered as
learners; and gained a new sense of community. The results showed
that Internet access enabled "powerful emotional and psychological
transformations" on the part of the participants.
Educational Software Use By
Educationally Disadvantaged Students
During the past two years, my colleagues at BBN have used the
computer program RelLab (for Relativity Laboratory) to teach the
concepts of relative motion to educationally disadvantaged
inner-city high school students in Boston (Horwitz, 1995). The
program enables users to construct and run relativity "thought
experiments." The inner-city students worked
OCR for page 397
Page 397
on the same sequence of activities as the educationally
advantaged suburban students with whom we worked earlier-and with
almost identical results. They advanced steadily from simple
Galilean (low-speed) relativity problems to complex ones involving
the frame independence of the speed of light and its implications
for simultaneity, time dilation, and length contraction. Along the
way they exhibited the same frustrations and overcame the same
obstacles as the suburban students, and they progressed at about
the same rate. Moreover, they showed as much interest and had as
many breakthroughs as their advantaged peers. The only real
differences were that the inner-city students were considerably
less verbally communicative than the suburban students and almost
all were severely educationally deprived. This was particularly
evident in the case of math. Their knowledge of the decimal system
was spotty and unreliable, and they had great difficulty graphing
data. Despite their competence on the computer, they were hopeless
at pencil-and-paper tasks with the same material, particularly
written tests. Based on posttest interviews of the students, it
appeared that their poor performance was due not only to knowledge
gaps but also to significant deficits in reading comprehension.
During the past year BBN researchers have also been using
another educational computer program, GenScope, to teach the
concepts of genetics to suburban and inner-city students in high
school biology classes. GenScope enables users to explore and
experiment with genetics models. It has proved engaging to both
male and female students of widely differing ages, backgrounds, and
ability levels. The most interesting results have come from the
inner-city population, where students who had learned very little
from a conventionally taught biology class learned a great deal
about complex genetic processes working with GenScope and used
their new knowledge effectively to accomplish a variety of
analytical and constructive tasks. Moreover, they remembered what
they had learned, and some were able to give accurate and detailed
explanations weeks later in interviews conducted away from the
computer. However, we found the same results as those noted above
with RelLab-though the inner-city students often learned to do
sophisticated genetics on the computer, their knowledge rarely
transferred to paper-and-pencil tests. They were unable to express
in writing the complex knowledge they exhibited on the computer.
Given the software tools and support, educationally disadvantaged
students acquired and applied complex reasoning, but their
knowledge was locked in. They lacked the basic reading
comprehension and communications skills that are fundamental for
success in education and skilled occupations. The challenge is to
help students like these acquire the verbal literacy that will
enable them to participate more fully as citizens in the knowledge
society.
OCR for page 398
Page 398
Research Issues
The positive outcomes shown in the study of Internet use by
low-income families are indeed impressive and encouraging. This
work cries out for replication and suggests new questions for
further research. What are the key requirements for making such
Internet use productive? What skills and motivation must users
possess to begin with? What kind of training and ongoing support
are required? How much on-site help is required, and how practical
and effective is on-line mentoring in augmenting initial training?
Can the facilities, training, and support be provided not only or
not primarily in the home but also in social settings like
community organizations, churches, local schools (after-school
programs), and particularly libraries-places with expert human
information resources?
The results from the use of sophisticated learning technology by
inner-city high school students-their success in acquiring and
using complex knowledge contrasting strongly with their difficulty
in communicating that knowledge-are more problematic and troubling
and raise another set of research questions. Can high school
students who fail to acquire competence in reading, writing, and
communications skills overcome these deficits later? Are these
difficulties remediable by tutoring? Or is there a literacy
learning barrier for English (analogous to the putative
foreign-language age block) that makes the acquisition of fluency
in reading and writing English enormously difficult after a certain
age or developmental level?
These questions provide a rich source of cognitive, educational,
ethnographic, and technological research investigations, from
intensive small-scale studies of individual development to
long-term longitudinal programs involving large populations. Rather
than providing brief summaries of possible research projects across
this wide spectrum, I will focus on a single project that is
motivated by both of the preliminary studies discussed above and
that has important implications for advancing the goal of universal
citizen participation and empowerment through educational
networking.
The productive use of the Internet by low-income citizens would
not have been possible if they did not bring to these activities a
fairly high level of functional literacy. The stunted development
of literacy skills among the inner-city high school students
severely limited their social participation and learning
opportunities, despite their inherent intellectual abilities.
Effective utilization of the NII, no matter how rich its user
interface and information resources, will require that its users
bring a modicum of literacy. But our nation has an enormous number
of citizens who do not meet that test. I describe a research
project that seeks to address, through the use of new interface
technology and associated instructional
OCR for page 399
Page 399
activities, a most severe literacy problem-that of American
adults who are functionally illiterate.
The Adult Illiteracy Problem
The development of a technologically literate work force is a
national imperative for the United States in the face of increasing
international competition. Major federal initiatives such as the
School-to-Work Opportunities program and the Advanced Technology
Education program have been created to address this critical need.
However, these efforts will be severely hampered if the country's
massive adult illiteracy problem is not solved. In confronting our
formidable economic challenges we need to recognize that the most
central and essential prerequisite for virtually all types of new
jobs is functional literacy-the ability to read and comprehend text
and to use various forms of word processing and other
communications software.
Workplace literacy surveys indicate that over 90 percent of
current occupations require reading. A recent study by the U.S.
Department of Labor and the American Society of Training and
Development found that, on average, workers spend more than 1.5
hours each day reading such materials as forms, charts, manuals,
electronic display screens, and general literature. Further, these
requirements are increasing: only 4 percent of new jobs can be
filled by people with the lowest levels of literacy, as compared
with 9 percent of existing jobs. According to the study, even the
one-third of American workers who perform production and service
delivery will need to read at an eighth-grade level. The remaining
two-thirds will need to read at postsecondary and higher levels.
Levels of literacy that were once acceptable will be marginal by
the year 2000.
However, the stark reality is that the number of adults in the
U.S. population with unacceptable levels of literacy is enormous.
Already, more than one out of five adult Americans are functionally
illiterate, and their ranks are swelling by about 2.3 million
persons each year. Nearly 40 percent of minority youth and 30
percent of semiskilled and unskilled workers are illiterate.
Illiteracy costs the United States over $225 billion annually in
corporate retraining, lost competitiveness, and industrial
accidents. The implication is clear: our goal of providing a modern
competitive work force hinges very directly on our ability to
achieve a massive improvement in adult functional literacy during
the next decade. This cannot be accomplished through the use of
human teaching alone. There simply are not enough reading
instructors. Their teaching must be augmented by the creation and
widespread application of an effective technology for automating
literacy tutoring.
OCR for page 400
Page 400
Although for a small fraction of illiterates the ability to read
is impeded because of neurological problems, and for others there
are learning difficulties that are not associated with sensory or
motor problems, the primary cause of illiteracy among Americans is
a failure to learn to read. Learning to read requires time and
practice. Research indicates that once the basics of learning to
read are in place, a grade-level gain in reading ability takes
approximately 100 hours of engaged literacy training time. Further,
at beginning levels of reading, individual feedback, motivation,
and guidance are critical. For most adult illiterates a major
obstacle to effective reading development lies in two simple
facts-the human resources do not exist to provide the teaching
support that is needed, and there is no way to adequately increase
their number to provide such support during the next several years.
A sufficient force of trained professionals and paraprofessionals
at the level of expertise required cannot be developed, even with a
massive injection of funding. The only option is the effective
introduction of appropriate technology.
Research Project On Adult
Illiteracy
My thesis is that one technology in particular-computer-based
speech recognition-provides the central and essential capability
required for launching a significant attack on a major segment of
the adult illiteracy problem. The reason for this view is
straightforward. Many adults who have serious reading difficulties
can speak English intelligibly, even when their speech is in a
dialect other than standard U.S. dialects. We at BBN plan to begin
from their strength-their ability to speak English. They
know how easy it is to talk and how hard it is for them to read.
They do not know that their own speech can be transformed into
text, that there is a direct, albeit complex, correspondence
between the two forms of language. Even though they cannot
initially read the text that is produced, they know that it is
their own, that they created it from their own speech. So in a very
real sense they know what the text "says." Further, they can repeat
the utterance, either in its entirety or partially, a word or
phrase at a time. In doing so, they begin to get a handle on the
translation problem through a procedure they substantially control,
in ways and at rates they find comfortable.
The starting point in this instructional strategy is to have the
computer generate a speech utterance. The trainee is then asked to
repeat the utterance, to "say the same thing." The speech
recognition system prints the correct text corresponding to the
utterance, but it remembers the trainee's articulations of the
constituent words and phrases, including those not spoken in
standard English. So the system develops a knowledge of the
trainee's idiosyncratic pronunciations and the text of the
words
OCR for page 401
Page 401
and phrases to which they correspond. This information will be
used to generate the correct English text in activities where the
trainee takes the initiative in generating the utterances.
This training strategy is distinctly different from traditional
approaches based on giving trainees a reading task from the start.
Instead, reading skills are developed in a more natural and gradual
way from speaking and listening skills that, to begin with, are a
great deal stronger. The transition from the scaffolding of
speech-driven pattern matching to text-driven decoding and
interpretation is not simple, but the confidence inspired in the
trainee by being able to begin to make sense of text should
substantially help in confronting and bridging this gap. In later
phases of the training, students will be engaged more in practice
activities to extend their vocabulary and to apply reading in
contexts of use through task-oriented activities. The training
sessions will involve the use of workplace literacy materials,
including interactive simulations, games, and extended stories,
gradually increasing in the level of challenge and complexity. The
functional literacy tasks will be based on those defined by the
National Adult Literacy Survey, particularly those in the first
three levels (e.g., interpreting bus or airline schedules;
following written and illustrated instructions from a manual or
display for such tasks as repairing a paper jam in a copying
machine). Interactive computer simulations of such tasks will
involve the development of procedural reasoning and problem-solving
skills in addition to reading comprehension skills.
This approach requires continuous, real-time, high-accuracy,
dialect-sensitive speech recognition. With the introduction of such
new and powerful facilities, the development of a literacy trainer
incorporating two-way interactive speech technology has become
feasible. Earlier speech recognition systems were not capable of
recognizing with high accuracy, naturally spoken utterances
involving large vocabularies and complex grammars. Moreover, they
were not "speaker independent" in that their accuracy for any
speaker was highly dependent on whether the system was tuned to the
speech characteristics of that speaker, through a tedious,
time-consuming "training" process. All this has changed with the
introduction of systems such as the BBN HARK recognizer. Given the
vocabulary and grammar associated with a student session, HARK can
recognize utterances with a word accuracy of around 95 percent.
Furthermore, the system can readily be configured to recognize
speakers with Haitian, Cambodian, Hispanic, and other common
dialects.
This experimental research and development project would be
conducted jointly by industry and university researchers. I
envisage a 10-year program to implement a comprehensive yet
inexpensive training system, demonstrate its instructional
capabilities, and evaluate its learning benefits with a
representative group of trainees chosen to exemplify a
OCR for page 402
Page 402
wide range of literacy problems. A two-year small-scale pilot to
develop a prototype system and demonstrate the instructional
approach might constitute the first stage of the project. It would
endeavor to show how speech-mediated literacy training technology
can effectively be used to make major inroads on our national
illiteracy problem. Speech scientists, instructional researchers,
and software developers are keenly interested in participating in
the proposed work. The project would have the long-term goal of
substantially reducing the adult illiteracy problem in the United
States and opening the way for fuller participation of all citizens
in the knowledge society.
References
Bier, Melinda. 1996. Personal Empowerment
in the Study of Home Internet Use by Low-Income Families. World
Wide Web article,
http://www.educ.ksu.edu/projects/jrce/v1/Bier/article.html.
Horwitz, Paul. 1995. Electronic
communication, Department of Education discussion group on a
National Technology Plan, March 20, Kirk Winters, moderator.
Archived in inet.ed.gov under NTPlan.
National Telecommunications and
Information Administration. 1995. Falling Through the Net: A
Survey of the "Have-Nots" in Rural and Urban America.
Washington, D.C.: U.S. Department of Commerce.
OCR for page 403
Page 403
Electronic Access To Services For
Low-Income Populations
Adam Porter
University of Maryland
Introduction
Much of the Internet's popular appeal stems from its potential
to provide a global virtual marketplace for goods and services.
Already high-profile projects are under way to develop video on
demand, virtual malls, and massive digital libraries. In addition
to these glamorous products, more mundane government services will
some day be delivered (in whole or in part) via the Internet. These
services might include Medicare, welfare and unemployment,
immigration, and job placement and training assistance. Since many
of these services exist to serve low-income populations, it is
important to determine whether low-income users differ from other
Internet users and, if so, what implications this has on user
interface design.
For the past several years I have been working with a nonprofit
group called Raising Hispanic Academic Achievement (RHAA). RHAA is
located in the Washington, D.C., metropolitan area and provides
academic tutoring and mentoring to several hundred Latino children
and their parents. Over 70 percent of these children come from
families whose annual income is below $20,000 and many receive
government services such as those described above.
Surveying The User Community
I conducted an informal survey of this user community to help
understand its characteristics and needs. This is not a scientific
sample, and it is clearly incomplete. Nevertheless, it provides
some insight into issues to consider when designing interfaces for
these types of government services. The survey identified several
characteristics that might not be found in other user
communities:
•
Non-English
speaking. Many respondents are first- or
second-generation Americans. Consequently, many do not speak
English at all or have limited proficiency.
OCR for page 404
Page 404
•
Limited educational
background. Many of the respondents have
not completed high school. Moreover, they have had little exposure
to computers and computer programming.
•
Limited literacy. Since many of the respondents do not speak English and
have limited education, functional illiteracy in English is high.
Furthermore, 20 percent (estimated) of the parents I surveyed are
also illiterate in Spanish.
•
Limited access to
computers. Few of the respondents have
computers in their homes. Those who do tend to have lower-end
machines with limited storage and printing capabilities. Almost all
have phone or pager service. Some of those without home computers
have access to the Internet through schools and/or public
libraries.
Considerations For User Interface
Design
Based on this survey I have identified several issues and
concerns that should be considered when developing user interfaces.
Many of these, of course, will apply to other communities as well.
I have grouped these issues into three categories: (1) computer
literacy, (2) language literacy, and (3) limited computer
resources.
Computer Literacy
•
Simple predictable
interfaces. Users often have to formulate
queries to search large information spaces. Frequently, they must
write these queries using SQL or logic programming languages. Since
many of these users have no programming experience, this type of
interface leads to errors and should be avoided when possible.
Also, interfaces should be predictable, not changing from use to
use, since this can lead to confusion.
•
Rapid, incremental, and
reversible control. Users should be
encouraged to navigate large information spaces quickly and easily.
Rather than programming queries, users should be able to visually
define queries and then refine them rather than recomputing from
scratch. Also, each operation should be reversible.
•
Training. One research area that should be explored is the
development of novice versus expert interfaces. To support novice
users better, error recovery and prevention schemes will be
important, as will further study into architectures for "help"
systems.
Language Literacy
•
Direct
manipulation. Since even experts make
frustrating typing mistakes when using textual interfaces, the
illiterate will find these interfaces
OCR for page 405
Page 405
unusable. Although direct
manipulation interfaces alleviate some of these problems, there is
still a strong need for more effective visual presentation
approaches.
•
Audio support. For people who cannot read or whose language is
unwritten (e.g., certain Creole languages), audio support may be
necessary. In these cases, text-to-speech systems may provide
low-cost conversion of written content.
•
Culturally appropriate
presentation. Different languages are
presented and organized differently. They have different accent
marks, character sets, and orientations in which words are read and
written. (This problem also appears in commercial software
development where products must be internationalized.) Rather than
expending scarce resources to translate content, reformat output,
and redevelop character-processing code, some research should be
devoted to low-cost translation systems, flexible software
architectures, reconfigurable browsers, and so forth.
Limited Computer Resources
•
Presentation on low-bandwidth
devices. Many people do not have
computers in their homes and have limited access elsewhere. As
interface developers we cannot assume that everyone is using
Netscape. We need to explore methods for low-cost conversion of
content for different devices (e.g., phones, beepers). For example,
Lucent Technologies has developed a language that allows HTML
documents to be presented in one way for a phone and in another way
for a standard browser.
•
Users without fixed Internet
addresses. Low-income people may have
computer access only through public channels, such as libraries and
schools, rather than at home. Notification services that assume a
fixed or a forwarding address will be inadequate. Some research
issues to consider are secure identification mechanisms and
flexible locator services.
•
Long-running
services. Providing services may require
may steps, with human intervention at several points. Therefore,
users may need to suspend long-running services and resume them
later (possibly from a different physical location, with different
computer resources, user names, etc.). Consequently, architectures
need to be developed for incremental interaction and interfaces
that summarize interaction histories.
Summary
This paper explores the design of user interfaces for systems
that provide electronic access to government services. Since these
systems exist to serve low-income populations, I conducted an
informal survey of one low-income population to understand their
needs. The survey indicated
OCR for page 406
Page 406
that low-income populations will have many novice computer
users, require support for languages other than English, and have
limited access to computer resources. Since redeveloping content
for every user community would be prohibitively expensive, this
situation presents a wide variety of research challenges. Three of
the most interesting topics are (1) visual presentation and
searching of large information spaces, (2) low-cost translation and
conversion of content, and (3) flexible software architectures and
interaction patterns.
OCR for page 407
Page 407
Access For People With
Disabilities
Larry Goldberg
WGBH Educational Foundation
Introduction
The barriers that prevent persons with limited sight, hearing,
or mobility from gaining access to the fruits of the emerging
national information infrastructure (NII) are not all that
different from the barriers that face all the other millions of
Americans who have, voluntarily or not, opted out of participation.
Limited bandwidth, limited technological resources, limited
technological facility, limited income, limited time, and limited
interest are the restraints bifurcating our society. This division
is self-perpetuating in that without interventions like the
promotion of every-citizen interfaces the gap will widen and feed
on itself.
For people with disabilities, input/output issues are the core
concerns. Disabled people share with their able-bodied colleagues
concerns about the complexity of systems, training needs, flexible
and intelligent interfaces, and so forth. But without the ability
to input commands and data and receive appropriate output, the rest
of the issues are irrelevant.
Today's Media
For the more basic, one-way components of the NII (i.e.,
television in its broadcast, cable, satellite, video-on-demand
formats), provision of access to the content flowing over these
pipelines continues to be an after-the-fact retrofit. That is, upon
finalization of programming, closed captions or video descriptions
are added. In the case of the former, caption data are encoded into
the vertical blanking interval (VBI) of the television signal and
delivered and displayed to the end user via a set-top box or
built-in decoder chip. In the case of the latter, an additional
audio track is added to the master video and delivered to the
end-user via an auxiliary audio channel (the Second Audio Program)
on stereo televisions and video cassette recorders.
Though widely utilized for decades (closed captioning) and years
(descriptions), these technologies are by no means 100 percent
reliable, owing to both system errors and human errors. Research
into fail-safe
OCR for page 408
Page 408
mechanisms for delivery of these data would provide the benefit
of consumer confidence and programmer satisfaction.
In addition, the distribution plants of the major broadcasters
and cable system operators often either cannot traffic the extra
audio channel for video description or will strip those
descriptions prior to delivery to the home. Preliminary research
into digitization of the auxiliary audio channel and encoding into
the VBI (like in the captioning process) has been conducted with
promising but inconclusive results. The added question requiring an
extra user appliance for decoding VBI-encoded audio needs to be
examined. Research may determine means of adding such capability
into today's or next-generation television systems.
Production of access services also is time and labor intensive.
Though trivial in terms of the overall cost of mass media
production, budgets for access services are carefully scrutinized.
This has become an even greater concern as the sources of
information and entertainment continue to grow. As we approach the
ability for every citizen to become a programmer/distributor, the
need to facilitate and lower the cost of producing and adding
access adjuncts or alternate means of output (text and audio)
becomes more dire.
Potential research in this area needs to focus on the
utilization of speech-to-text and text-to-speech technologies.
Neither has yet been developed to the point of aiding in the
production or distribution process of either captions or
descriptions, but both show tremendous promise. Today, most of the
burden for providing captioning on live programming falls on a
extremely small group of highly skilled ''stenocaptioners" who can
outgun any speech recognition engine available today.
Captioning has clearly proven itself to be a service that serves
populations far beyond the originally intended audience. Early
optimistic explorations into the use of captioned television for
early and remedial readers has been followed by exploitation of
captioning for students of English as a second language.
Computer manufacturers are now building caption displays into
their "personal computer/televisions" with the added capability of
downloading caption data for archiving and indexing purposes. And
researchers in the field of digital video storage have long been
excited about the use of captions as indices for large-scale video
databases for research, archiving, and production uses.
Perhaps more trivial but ever more evident, captioned television
has become widely used in environments that require an alternative
to audio output, such as health clubs, bars, libraries, and for
late-night television watching.
Video description uses for additional populations have just
begun to be explored and would certainly benefit from significant
research projects.
OCR for page 409
Page 409
Preliminary work has begun to determine the potential of
description to assist students and adults with learning
disabilities, especially attention deficit disorder. Additional
similar benefits may be discovered through research into video
description's impact on information retrieval where masses of data
and information clutter could be reduced through "descriptive
guides."
Embedded description has also been suggested as a potentially
useful way of indexing large visual databases if coupled with
intelligent speech recognition. Little or no research has yet been
conducted in this area, but that may point the way to another
repurposing of an access technology.
New Media
In the worlds of nonlinear and digital multimedia, interactive
media, and advanced television, many problems remain to be solved.
Standards are either evolving or have not yet been considered for
incorporation of captioning and description in all of these venues.
Early efforts at creating such standards ignored the ancillary uses
of these technologies and threatened to lock in bad designs.
For example, the ability to display text in the emerging DVD
(Digital Video Disk or Digital Versatile Disk) format has been
designed with language translation as the primary purpose for text
display. Since the DVD format has mostly been designed in Japan,
the text is being created through bit-mapped graphics files, not
ASCII text as in broadcast closed captioning. This has resulted in
many unforeseen design flaws: the inability of a DVD player to
address the closed-captioning circuitry in television sets, the
inability to flexibly change fonts or type sizes, and the inability
to use the text as a search engine.
These decisions are made as new media products rush to market,
with little attention paid to the best-possible design for the
built-in access services and how they can be configured for
universal applicability. Similar concerns are being directed toward
the designs of the alternate audio features of advanced television
and DVD.
As great a concern has been expressed about the "appliances"
people will be using to access the Internet or other digital
information and entertainment systems. With personal computers the
ability to add software or hardware to assist those with special
needs is relatively straightforward (though not without significant
programming and development efforts). But if every citizen will be
accessing the new services via "thin clients," smart phones,
set-top boxes, or low-end, low-cost browsing boxes attached to
television sets, the ability to add such services as speech
synthesis or output to refreshable Braille or large-print displays
is problematic.
OCR for page 410
Page 410
When digital sound files are sent over the Internet and received
via these boxes, how will the boxes turn the sound into text for
deaf people?
Research into the ability to incorporate these alternate output
modes into inexpensive information appliances can result in more
accessible low-end browsers for every citizen, including blind and
deaf people. Incorporation of new access standards into digital
media formats (such as QuickTime, Active Video, and RealAudio) will
facilitate the ability of low-end browsers to display incorporated
access technologies. All of these research challenges can be
addressed in the short term (one to five years), except perhaps for
the need for a fully capable speech recognition technology, which,
after 20 or more years of effort, still requires many years to
approach the speed, accuracy, and other capabilities needed for use
by deaf people.
It has been the experience of those in the world of media access
that these design and research challenges are not complex
(certainly not as complex as the creation of the new media
themselves). When focused attention and resources are applied to
the problems, solutions are readily discovered, especially when
approached by consortia of public and private practitioners and
researchers. Early awareness and design-from-the-blueprint-stage
thinking obviate the need for expensive and inefficient retrofits
that are resisted by producers and consumers alike.
Suggested Uniform Resource
Locators
http://www.wgbh.org/caption
http://www.wgbh.org/dvs
http://www.wgbh.org/ncam
http://trace.wisc.edu
Representative terms from entire chapter:
adult illiteracy
Items in cart [0]
