Volume III Number 1, March 1996

Ease Of Use And Maintenance: A Note On Software Design

Richard Plant
Northern General Hospital
Sheffield, UK
Andrew Rostron
University of Hull

Increasingly powerful hardware has prompted commercial software developers to concentrate not on highly efficient code but instead on usability. With much modern software sharing the same look and feel, a commercial edge can only be gained by increased usability leading to improved productivity. In the commercial world, usability ratings by NSTL (National Software Testing Labs) and similar institutions can affect sales significantly, in the same way that a poor review might prematurely end a play's run in the theatre.

In the field of voice output communication aids (VOCAs), client usability and high productivity (in terms of speech output rates) have always been priorities. However, much of this usability has typically centered around users themselves and not the rehabilitation worker, or health care professional, who might participate in developing the range of materials and experiences (maintenance) offered by such an aid. In fact, even in Beukelman and Ansel's (1995) enumeration of research priorities in augmentative and alternative communication, no direct mention is made about ease of maintenance of VOCAs by those who have to structure the contents of these devices. As VOCAs take account of new technologies and become increasingly advanced, they may remain highly accessible to users, but increasingly complex for typical maintainers who have to develop material for them. For example, a DECTalk/symbol based aid might require the writing and maintenance of scripts. This process can be complex, tedious, and, in time-pressured service environments, counter-productive.

A simple analogy illustrates the point. At Christmas, a child may receive a toy which needs to be assembled by an adult. The child sees only the completed toy with which they are very happy, but does not appreciate the effort expended by the adult. If the toy were constructed from fewer parts which were less complex it could be constructed more rapidly. This would enable the adult to have a better understanding of the toy's workings in case they needed to fix it, to play with the child for longer, and to reduce potential stress associated with the construction. In a similar way VOCA can be difficult to understand from a maintainer's prospective, which may mean that maintenance times are increased and frustrations raised. Naive maintainers are likely to suffer most in this respect.

Concentration on usability, although highly desirable, can mean that maintainers may easily become disillusioned about a specific VOCA. Negative perceptions regarding VOCA maintenance are usually to the detriment of the user. On the other hand with good design a VOCA can be made to be rewarding and productive for both users and maintainers. If a device enables maintainers to build complex materials easily, and offers instant rewards, ultimately this is likely to be of benefit to the user.

With modern philosophies of software design such as object orientation, the maintenance of VOCA can be made simple and rewarding if the maintainer interface is correctly designed. It is suggested that maintainers' attitudes and abilities to use a given VOCA should be incorporated into the development cycle as is currently the case with users. With the employment of standard user interface questionnaires, such as the QUIS (Shneiderman, 1992), failings can be addressed and strengths consolidated during the development phase. In our own research, by following the principle that users and maintainers share the same good design needs, such concerns have been addressed in parallel. In this instance the needs of users and maintainers are considered of equal importance.

This ethos has produced a software based VOCA known as "Easy Speaker" which, from a user's perspective, seamlessly integrates photorealistic symbols, digitized speech, video clips, animation, MIDI music and the like. By making use of dynamic displays, the user can communicate vocally, in addition to using the device as a learning and recreational environment in its own right. By maintaining a simple user interface, demands on users have not increased to reflect these enhancements. Interaction remains simple but offers a far richer environment.

One might expect that an increased feature set necessarily increases the workload for the rehabilitation worker as it might in many other VOCAs. However, appropriate design means this is not the case. By using an object-orientated design philosophy, maintainers can drag scanned photorealistic images on to any of the dynamic screens the user can see. Speech can then be recorded to accompany these symbols at the touch of a button. By using graphical controls that mirror a Dictaphone, or video cassette recorder, caregivers are already familiar with the notion of pressing the red record button to record speech. Symbols that are incorrectly positioned can be moved at will by dragging and dropping elsewhere. Alternatively they can be dropped into a bin which appears at the bottom of the screen. Whatever the form of media, it can be dropped into place, ready for use.

In short, the object-orientated nature of the design successfully mirrors the rehabilitation worker's understanding of the real world. For example, when sticking pictures in a scrap book, you move them around on the page until you are happy with the layout and then glue them in place. In our software, you drag them around and then save the layout as opposed to the use of glue. By using such synergistic design, the rehabilitation worker can efficiently construct large vocabularies of digitised speech and other media represented by photorealistic symbols. Unlike other VOCAs there is no complex scripting language; for that matter, there is very little else to learn. In this instance the overall objective was to produce a system that could be mastered by those familiar with current graphical user interfaces in a matter of minutes and by computer novices after a few short training sessions. Our first experiences with this latest version of Easy Speaker suggest that some considerable progress has been made in achieving this aim, and so through good design the needs of both users and rehabilitation workers have been addressed. Previous experience with an earlier version of Easy Speaker, which did not have the range of facilities of the latest version, illustrated that some clients, including some with severe learning difficulties, found it had much to offer; Rostron, Plant and Herrman (1994), Rostron, Plant and Ward (1996), did not find it sufficiently user-friendly.

As a complement to object-orientated design, other facilitative features have been incorporated. These include the use of status bar messages that change as the caregiver moves the pointer over screen objects. As a complement to printed manuals, full context sensitive on-line help is provided, which can be searched and printed at will. An Automated Response Tracking System (ARTS) has also been incorporated that tracks utterance construction, timings, cursor efficiency, actual and estimated words per minute scores, etc. Such measures can be used in both assessment and performance monitoring, and more importantly on the basis of these the layout can be modified accordingly.

In sum, the view that both the user and the rehabilitation worker are of equal importance should be taken when developing any communication aid. By following this ethos we have produced a VOCA which is liked by users and maintainers alike. Just as in the commercial software industry, new VOCA may be differentiated on the basis of their usability. It must be remembered that attitudes and decisions made by caregivers and maintainers typically have a greater impact on VOCA users than does the design of the user interface per se. This begs the question of whether the maintenance interface is actually more important than the user interface? If the former is poorly designed this can only ultimately be to the detriment of the user.

The latest version of Easy Speaker is now available, (for PCs only), together with full documentation, via the Internet site - or
http://www.hull.ac.uk then Psychology


Beukelman D R and Ansel B M (1995). Research Priorities in
Augmentative and Alternative Communication . _Augmentative and
Alternative Communication_, 11, 131-134.

Rostron A B, Plant R and Hermann C (1994). Hypermedia for the
learning disabled: expensive luxury or useful tool? _Computer
and Education_, 22, 215-224.

Rostron A B, Plant R and Ward S (1996) _European Journal of
Disorders of Communication_ In press

Shneiderman B (1992). _Designing the User Interface:
strategies for effective human computer interaction._ (2nd ed.)
Reading, Massachusetts: Addison-Wesley.

Plant, R. & Rostron, A. (1996). Ease of use and maintenance: A note on software design. Information Technology and Disabilities E-Journal, 3(1).