Test your Site for Accessibility with Cynthia Says ™

A Proposal for a Research Agenda in Computer Human Interaction

William P. LaPlant, Jr.
Technology Research Staff,
Statistical Research Division, U.S. Census Bureau of the Census
Washington, DC 20233-9100
blaplant@census.gov; blaplant@minspring.com; 301-457-4887


After reviewing the factors which necessitate a revolution in human accessibility to intelligent technologies, this paper proposes a comprehensive approach to the reporting of past and current research, and the planning of future research in the area of the human computer interface (HCI).  A more rapid transfer of HCI research results from academia to industry can be accomplished by developing a framework for managing and reporting those results.  Industry benefits by having research data available faster and in a form more usable by product designers, engineers, and usability specialists. The Computer Human Interaction (CHI) research community benefits from having a tool that will support identifying the areas needing new research, replicating existing results, and providing an ongoing source of pedagogic material.


There has been a great deal of research on human performance at the Human Computer Interface (HCI).  Significant portions of  this conference and all of its predecessors have been devoted to reporting the results of that research or its subsequent use.  The results of this research are always reported in terms of the experimental designs and of the hypotheses being tested, and therefore may be most useful in developing subsequent, similar research. The problem is that the results of this research may not be quickly and easily applied in the engineering and product design processes.

At the same time, the need for quick and broad application of this research is becoming obvious. More and more of our lives are affected by computers and the Internet in more and more ways. Thus it is becoming essential for all of us to have full, effective access to computers and the Internet in order to be a full participants in our modern society.  But studies have shown that there is a substantial "digital divide" that prevents many of our fellow citizens, particularly those with disabilities, from such full participation.

In order to accelerate this essential "technology transfer," I propose that the Computer Human Interaction (CHI) research and user communities develop mechanisms to organize, normalize, and extend what we know about how human beings (both with and without disabilities) perform using various combinations of HCI input and output modes. This would involve research in a number of areas:

  • A taxonomy (or ontology) for human performance and technology at the HCI (the "problem space");
  • Measures and common reporting mechanisms for human performance at the HCI;
  • An exhaustive survey of all existing work at the HCI;
  • Research on multi-modal combinations, tasks and interfaces not yet done;
  • Development of one or more predictive models of human performance at the HCI,
  • Systematic verification of existing work in support of predictive model and experimental result verification, and
  • Creation of a National (or International) knowledge base on CHI research results.


There are trends in information technology and society that necessitate a revolution in HCI technology transfer. More and more of our lives are affected by Information Technology (IT) and the Internet in more and more ways. It is rapidly becoming essential for everyone to have full access to our national information infrastructure (NII)in order to be a full participant in our modern society.

Internet Domain survey Host Count. D

Trends in Technology:

As you can see from the diagram[1], the Internet is growing at an exponential rate[2]. And the amount of data and information available to anyone with access to Internet is growing faster than the Internet itself[3].  Technology itself is improving an speed and capacity. As predicted by Moore’s Law[4], computer processors, the underlying technology of the information age, have been increasing in size exponentially.  Moore’s Law has apparently held true since Moore first observed the trend in 1965.  Variants Moore’s Law appear to apply (with different constants) to the ever increasing speed of communications, and to the increasing speed and capacity of on-line permanent storage.  All of this fuels the explosive growth of the Internet and the World Wide Web. This in turn fuels a trend toward being able to use the Internet to transact most of the activities of daily living. For example, there is the growth of "e-Government" at all levels[5]. "E-Commerce" is growing in popularity even faster[6].

Technology is also permeating every aspect of our lives.  Indeed, the non-PC use of the Internet is growing far faster than the PC-based Internet. 

Internet Growth ChartD

This diagram from the 2001 Technology Forecast issue of IEEE Spectrum[7] shows this trend clearly.

The response of industry has been to develop new ways for intelligent devices to interact together and with human users. As a result there is an ever increasing amount of intelligence in the devices we use in our daily lives. Here is a (non-exhaustive) list of the industry groups and technologies supporting the further exploitation of that trend:


At the same time the population that clearly needs assistance in taking advantage of this technology is growing. The population in the U.S.[8] (and in the developed countries of the world[9]) is aging.  Further, more people are surviving with congenital disabilities, from disabling diseases, and from disabling traumatic injuries, so that the population with disabilities that is available for the workplace is increasing[10].

Digital Divide Statistics D

The Digital Divide:

In its 1997 book, More than Screen Deep; Toward Every-Citizen Interfaces to the Nation’s Information Infrastructure[11], the National Research Council (NRC) challenged industry and government to provide universal access for every citizen to the national information infrastructure (NII). The reason is obvious.  A number of studies[12] show, not surprisingly, that there is a substantial "digital divide" that prevents many of our fellow citizens, particularly those with disabilities, from such full participation. This chart[13], reporting computer use and Internet access by disability status based on 1998 data, shows that people with disabilities are about half as likely to have access to a computer at home and almost a quarter as likely to have access to the Internet, as people with no disability.

There is a growing need for new approaches to ensuring that everyone can use these technologies.  In 1998, both the U.S. Congress[14] and the President[15] provided guidance to federal agencies to ensure that people with disabilities are provided comparable access as the world move toward universal pervasive computing and inter-connectivity.  Regulations[16] have been issued, requiring federal agencies to acquire only electronic and information technology that can support comparable access to data and information by people with disabilities (both federal employees and members of the public using or providing data) to that provided to people with no disabilities. The need for various kinds of standards to support universal access has also been recognized by industry and professional organizations. In 1997, the Human Factors and Ergonomics Society submitted a requirement to the American National Standards Institute (ANSI) Information Infrastructure Standards Panel[17] for "standards … for a wide variety of software and hardware technologies to ensure accessibility of the NII to users who have disabilities."

Research Agendas:

There have been a number of research agendas proposed for the CHI community over the years. Almost all address either universal access, accessibility issues, or universal design as a motivating factor. 

In a recent paper[18] in Communications of the ACM, Ben Shneiderman identified a number of key areas for future research to support narrowing the gap.   Shneiderman proposes a research agenda that concentrates on "three universal usability challenges: technology variety, user diversity, and gaps in user knowledge."

Research agendas for HCI have been proposed over the last several years. See http://www.acm.org/pubs/articles/journals/surveys/1996-28-4es/a146-strong/a146-strong.html, http://www.acm.org/pubs/citations/journals/interactions/1995-2-1/p69-strong/,

These have been tightly tied to the requirements of pedagogy of HCI.


Others have addressed the social implications of HCI research and practice:

http://www.acm.org/sigchi/chi97/proceedings/paper/mm1.htm, or the research work of a particular research laboratory. For example, see: http://www.acm.org/pubs/articles/proceedings/chi/97243/p283-williges/p283-williges.pdf, http://www.acm.org/pubs/articles/proceedings/chi/67449/p51-moran/p51-moran.pdf, http://www.acm.org/pubs/citations/proceedings/soft/317498/p176-consortium/#abstract, http://www.acm.org/pubs/articles/proceedings/chi/97243/p85-harrison/p85-harrison.pdf,


I propose the development of a knowledge base in the domain of Computer Human Interaction in the broadest sense, which I will refer to here as the "Computer Human Interaction Knowledge Base (CHIKB)." The ultimate purpose of such a facility, from the perspective of meeting the needs and preferences of people with disabilities, is to provide the electronics and information technology (E&IT) industries with a tool needed to support true universal design for intelligent systems.  The result will be the near elimination of a need to treat people with disabilities differently from the general population in order to ensure comparable access to intelligent devices and to data and information essential to modern life. Development of the CHIKB will take a concerted commitment by Government, Academia, and the E&IT industries to reach this goal.  What are the steps need to create the CHIKB?

  1. Agree on HCI Research Ontologies or Taxonomies: First, both HCI researchers and users of HCI research results must agree on one or more ontologies and taxonomies that among them exhaustively cover HCI research that has been done in the past, is currently planned or underway, or may be done in the future.
  2. Develop Consensus Measures and Reporting of Results: Next, research results must be reported in a way that allows results to be most broadly useful.  This will require broad agreement on the methods and metrics used to determine human performance at the computer human interface.
  3. Record the Results of HCI Research Already Completed: This step involves surveying the entire body of past HCI research; determining how the results would be described in the HCI ontology; if possible, transforming the reported results or experimental data to the appropriate measure; and entering the normalized results in the CHIKB.
  4. Develop One or More Statistical or Empirical Models: With enough normalized data, it may be possible to develop predictive models using well known statistical or empirical modeling methods.

Benefits of the CHIKB:

Obviously to both academia and industry must participate fully for such an effort to be viable.  In order to make such an extraordinary commitment there must be extraordinary benefits for all involved.  Here are some of the benefits both communities and the public can expect as CHIKB is implemented:

  • Research and Pedagogy:
    1. As soon as the CHIKB is established, it will establish a method of rapid, peer-reviewed publication of experimental results.
    2. The models will provide a mechanism for determining the most fruitful areas for experimental verification.  If in the process of verification the experimental results are verified, the experimental methodologies will be confirmed and confidence in the model will have been strengthened. If the results are not verified, such work will lead either to revisions of the models, result in refinement of experimental methods or metrics, the discovery of other artifacts.
    3. As the CHIKB nears completion, it can provide an ever more reliable way of determining of "additions to knowledge." If no research is reported in a given area covered by the knowledge base, either there is a new area being explored, or a new way of considering the results has been developed.  Either is of value.
  • Use by Industry:
    1. As soon as the first results are reported in the CHIKB, because of the use of the "industry" concepts (in the ontology/taxonomy), normalized metrics, and common industry reporting format, industry will have rapid access to HCI research results in a way that can be quickly applied.
    2. As the predictive models are developed and verified, industry will have an ever more reliable tool for predicting human performance in intelligent systems in new situations (new technologies or novel applications).
  • All Users:
    1. Everyone can expect to be able to access intelligent devices in their environment in ways that are most effective and convenient for the task they are engaged in.  The days of the flashing VCR clock will be forever banished.
    2. People with special needs in making use of intelligent devices will no longer need extraordinary support to do it.  They may have fewer options than the general population, but access on an equal basis to those devices that are rapidly becoming a necessity for successful daily living and full employment will not only be possible, but routine.
    3. Government will no longer have to be in the business of ensuring that people with disabilities have full and comparable access to the data, information, and technology that increasingly enrich our lives.


Ontologies or Taxonomies:

The science supporting the development and use of ontologies is mature. See http://dlib.computer.org/tk/books/tk1998/pdf/k0513.pdf,

http://www.acm.org/class/1998/overview.html, http://www.isi.edu/natural-language/resources/shared-ontologies.ps, http://www.isi.edu/natural-language/resources/principles-of-il.ps,

Measures and Common Reporting Mechanisms:

Measures: http://computer.org/proceedings/hics/8341/83410182abs.htm, http://computer.org/proceedings/hics/8341/83410182abs.htm,

Reporting: http://www.nist.gov/iusr, http://zing.ncsl.nist.gov/iusr/documents/WhitePaper.html,

Surveys of Past Research:

Human Performance Research:

Develop Predictive Models of Human Performance:

http://www.acm.org/pubs/articles/journals/tochi/2000-7-2/p197-sutcliffe/p197-sutcliffe.pdf, http://www.acm.org/pubs/articles/journals/tochi/2000-7-2/p222-barnard/p222-barnard.pdf,

Systematic Verification:

Creation of a National (or International) Knowledge Base:

Obvious examples of such knowledge bases abound.  EPA maintains one on environmental research. NASA maintains on one on space research. There are two such systems for the International Genome project, one under public sponsorship and the second privately managed and maintained.

[4] (See: http://www.intel.com/intel/museum/25anniv/hof/moore.htm)  "Each new chip contained roughly twice as much capacity as its predecessor, and each chip was released within 18-24 months of the previous chip."

[7] See Perry, Tekla S., "Service Takes Over in the Networked World," IEEE Spectrum, Jan. 2001, pg. 102.

[12] For example, see both "Falling Through the Net" and "Falling Through the Net II,"  (http://www.ntia.doc.gov/ntiahome/fallingthru.html and http://www.ntia.doc.gov/ntiahome/net2/falling.html) and Disability Statistical Abstract 22: "Disability and the Digital Divide" (http://www.dsc.ucsf.edu/UCSF/pdf/ABSTRACT22.pdf).

[13] From Disability Statistical Abstract 22, pg 1.

[14] See Public Law 105-220, § 508, the Workforce Investment Act of 1998 amendments to the Rehabilitation Act of 1973. (http://www.usdoj.gov/crt/508/508law.html).

[15] See Executive Order 13078, "Increasing Employment of Adults with Disabilities."  (http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=1998_register&docid=fr18mr98-141.pdf).

[16] See 36 CFR Part 1194, "Electronic and Information Technology Accessibility Standards," at http://www.access-board.gov/sec508/508standards.htm and 48 CFR Parts 2, 7, 10, 11, 12, and 39, "Federal Acquisition Regulation; Electronic and Information Technology Accessibility," at http://www.access-board.gov/sec508/FARnotice.htm.

[17] See HFSE, 1997,  "IISP Need #161 – Human Computer (User) Interface: Enabling Accessibility for Users with Disabilities," (http://web.ansi.org/public/iisp/std_need/need161.html).

[18] See Shneiderman, Ben. "Universal Usability: Pushing Human-Computer Interaction Research To Empower Every Citizen." CACM, V43, N5 (May, 2000), Pg 84, ff.

This material was the basis of a presentation by the author to the ACM SIGCHI 2001 State of the Science Exchange (see http://www.trace.wisc.edu/docs/soschi2001/). The content will be included, in a modified form, in a report of the results of that event. The opinions and conclusions stated here are those of the author and do not represent an official position of the U.S. Census Bureau or any other agency of the U.S. Government.



You should also visit papersmart.net writing company to get your research papers and research proposals written by experts.