2001 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2001 Table of Contents


Patricia Walsh
E-mail: walsh@dots.physics.orst.edu 

Randy Lundquist
E-mail: randyl@physics.orst.edu 

John A. Gardner
E-mail: John.Gardner@orst.edu 

Science Access Project Oregon State University Department of Physics Corvallis, OR 97331-6507 http://dots.physics.orst.edu/ 

Steven Sahyun E-mail: sahyun@lclark.edu 
Department of Physics
Lewis and Clark College
Portland, OR


The Audio-Accessible Graphing Calculator is a self-voicing Windows application that has been under development and testing for some time by the Science Access Project. It includes the capabilities to: Beta testing was conducted in 2000, and in this paper we give an overview of the results of that beta test. In the presentation we will demonstrate how to use the calculator.

Summary of Beta Test Results

The Accessible Graphing Calculator (AGC) beta-test was run from May until mid-July, 2000. To qualify as a beta-tester, volunteers filled out a demographic survey form. Upon acceptance, volunteers then downloaded the AGC program from a Web site, or were sent copies of the program on CDROM.

There were 33 people who downloaded the AGC during the beta-test period. Some of the people who downloaded the program were individual testers, although there were several people responsible for installation and testing on many computers. Six people (some of whom downloaded the program) were sent CDROMs.

From data collected in the initial survey form, beta-testers ranged in experience from novice users to experienced programmers. Education ranged from high school students to Ph.Ds. Occupations generally fell into three categories: Adaptive Technology Specialists, students, and educators.

The locations of beta testers ranged world-wide with most residing in the US or Canada. Other countries that had testers were Germany, Egypt, Australia, and Palestine.

Testers heard about the beta test from demonstrations of the AGC at CSUN and other conferences (ICCHP), personal contacts, word of mouth (from special education teachers' ListServes and the Blind-Tech ListServ), and from the Science Access Project or National Science Foundation Web sites. Many (28) of the beta-testers were familiar with screen reading programs (such as Jaws, Window-Eyes, and Outspoken) although they did not necessarily use them. Only 8 were not familiar with screen readers.

Eight people reported that the program crashed during installation. It is well-known that many older commercial speech engines have fatal conflicts with the MS SAPI (speech engine) used by the AGC. When these old speech engines were uninstalled, the AGC program usually installed. No conflicts were found with the popular Eloquent speech engine used with many screen readers. We are aware of only one beta tester who removed the conflicting speech engines but was still unable to load the AGC program, and we have thus far been unable to find the reason for this person's difficulty.

Beta test volunteers were sent an e-mail survey about their experiences with the AGC. For testers who did not respond to the initial survey within a specified time frame, copies of the survey were e-mailed to them several times. The survey consisted of 22 questions (some of which had multiple parts) regarding their impressions about the program. There were 11 subjects who responded to the survey.

The AGC features that testers found particularly useful were: One feature that testers found annoying was that the help file is a standard Microsoft application that is used in the standard Microsoft-recommended way. This application is not self-voicing, so a screen reader is required to read it in audio. This annoyance was partially remedied in a later version by including a self-voicing help file accessed from the help menu, so users could choose their preferred format.

Other features that were criticized by some users were that the text was small at times, that the on-screen calculator could not be magnified, and that the program did not support braille embossers. It does support standard Windows printing however so output could be embossed with the new Tiger Tactile Graphics Embosser or printed on swell paper from which tactile copy can be made.

The calculator keypad was criticized by some as the buttons are in a standard visual layout. However, in general, beta testers found that the layout and implementation of features was good.

In comparison to other methods for accessing graphs, the AGC was generally considered the best. "Where the AGC is mostly effective is in its facility for giving the user a general sense of what the data looks like via the audio graphing function."

Ease of entering and editing equations, navigating between the AGC's features, and operating the data and graphing features, were rated on average between fairly easy and very easy. Actual scores were:
However it was noted that one must read the instructions first. One must also use the self-voicing capability for audio access. One beta tester insisted on using a screen reader and, not surprisingly, had a number of difficulties.

On average, beta testers rated the AGC quite practical, the number of features adequate for their needs (with the desire for more statistical functions noted), and the general layout good. On average, testers agreed that the benefits of using the AGC are worth the time and effort expended in learning to use the program and that the auditory graphs were a useful feature.

Nine of the 11 beta testers would recommend the AGC to others. The two who did not, cited concern for the time needed to learn and the difficulty of use with a screen reader.

Comments and Conclusions

The AGC program pleased most beta testers a great deal, although all would be pleased if it could be made easier to learn. They also noted several features that are unfortunate but unavoidable.

An on-screen keyboard cannot be magnified larger than the screen, so if used visually by people with low vision, a screen magnifier is required. We believe that a power user with low vision would be best served by using the hot keys for entering data and keep a screen magnifier on the calculator screen output. In a future release we shall attempt to provide an alternative to this calculator screen that can be magnified without a separate screen magnifier.

It was also noted that the program is not really usable with a screen reader. From the beginning, this program was intended to be universally accessible. In order to be friendly to sighted users one must adopt a two-dimensional screen layout for some features, and screen reader access immediately becomes a knotty problem. Any screen reader user who has tried to use an on-screen keyboard program, such as the calculators bundled with Windows, can appreciate the difficulty of using a screen reader where one really needs an information reader. We believe that we have solved this problem in the way it should be solved - by providing a true audio interface. Screen readers should just be put to sleep when focus is in the AGC program and other applications with full audio interfaces.

Finally, several beta testers were disappointed that AGC does not support braille embossers. Unfortunately, braille embossers do not have the capability of accepting WYSIWYG information, so it is not possible even in principle to emboss both the graph and its text labels. There are applications available that can be used to create graphs without labels to be embossed with standard braille embossers. Many observers feel that braille embosser resolution is just too poor for such graphs to be really useful. Two modern technologies, the Tiger Windows Embosser and Capsule paper, are now available that do permit fully-usable tactile copies to be made, and these work very well with AGC.

Although there are still improvements that we intend to make in future versions, we believe that the beta test results indicate that we have succeeded in making a program that fills a great need. At the time of writing this article, the AGC was not yet commercially available, but the authors are doing their best to introduce it commercially and hope that it is on the market by the time of the conference.


This research was supported in part by the National Science Foundation.

Go to previous article 
Go to next article 
Return to 2001 Table of Contents 
Return to Table of Proceedings

Reprinted with author(s) permission. Author(s) retain copyright.