John L. Bennett

Usability Engineering: Our Experience and Evolution

Publisher Summary This chapter discusses the users experience and evolution of usability engineering. Usability engineering starts with a commitment to action in the world. It seeks to capture user experience within a context situated in user work and in a form useful for engineering. Usability engineering provides operationally defined criteria so that usability objectives can be used to drive an efficient and productive engineering effort, and it can lead to production of systems that are experienced by users as usable and that serve as a basis for the next generation of systems. When a system is built and delivered to users, interaction with it would affect user experience and would shift the background against which users evaluate that system in comparison with other systems. Therefore, as systems are built that provide new functionality with new levels of usability, the expectations of users would shift so that the whole cycle should begin again.

Learning from user experience with groupware

Observers have identified a potential for major improvements in organizational productivity made possible through the use of personal computers serving as a means to link people into task-oriented teams. Our study offers an examination of how people are using personal computers for such electronic exchanges via networking. We interviewed 223 people who were using several “groupware” systems in a sample of 25 enterprises to see how they employ these software tools to support their group work. An explanation of our research design and a more complete discussion of the results can be found in Bullen and Bennett [BulNO].

User-oriented graphics systems for decision support in unstructured tasks

The quality of graphic interaction languages may be examined in terms of what the user sees at each interaction point, what he has to know in order to interpret what he sees, and what actions he can take at the interaction point. These guiding questions are of special importance in designing graphic facilities to be used by managers in solving unstructured problems found in business and industry. Three principles for design respond to the guiding questions and represent a partial synopsis of what we have learned in our work. I. Design the context presented to the user at each interaction point to suggest the actions which a user can take upon that context. II. When processes are indeterminate (e.g., in unstructured problems), focus design on particular displayable data representations (e.g. tables, graphs, charts) rather than on the processes in which representations may be used. III. Design the system to have an explicit framework so that a user working out a solution to an unstructured problem can take the graphics system structure for granted as he concentrates on discovering problem structure. These principles will continue to be developed, revised, and expanded in our Decision Support Systems project.

Learning HCI design: mentoring project groups in a course on human-computer interaction

We describe the role of mentors in an experimental course on human-computer interaction (HCI) taught in the Computer Science Department at Stanford University for the past two years. Students practice design within the course by collaborating in small groups on 12-week projects, in which they analyze a work environment, design and implement a prototype user interface, and evaluate the prototype with project clients. As part of our experiment in providing contact with the world of practical design, we invited people from local industry to serve as mentors for the student groups. These unpaid volunteers have been important adjuncts in guiding students as they developed their HCI projects. We discuss: the background for the projects; the role of mentors in the learning process; what is required of mentors and what benefits they get; how to support the mentoring process; and what we have learned from the experience so far. We consider mentors to be a critical part of the design project within the course, and we welcome a sharing of experience with other people who may have developed similar courses where mentors played an important role.

DEVELOPING A USER INTERFACE TECHNOLOGY FOR USE IN INDUSTRY

We are developing a user interface technology to address ease of learning and ease of use concerns on the user side of the interface during the design process. Modelling the user how-to-do-it knowledge required by a design is one step toward development of such a technology. We report on an evolving methodology that is intended to give developers early warning indications of potential usability problems that may arise from a set of design decisions.

Tools for building advanced user interfaces

System developers are noticing that their design decisions strongly affect computer usability. The design of the user interface has an important bearing on the knowledge users must have to accomplish work through the user-computer interface. Recognition of this fact is leading to the development of User Interface Management Systems (UIMSs). A UIMS is a design concept for separating the details of user interaction from the details of advanced applications. This paper shows how UIMS research and research into the representation of user process knowledge (i.e., user how-to-do-it skills) can help developers understand issues involving ease of learning and ease of use. This parallel progress in UIMS development and in user modeling makes it easier to build high-quality advanced user interfaces.

Facilitating effective HCI design meetings

Over several years we have participated as facilitators in many Human-Computer Interaction (HCI) design meetings. Our focus has been on developing team results needed to achieve user-centered design of software for computer systems. We describe frameworks for partnership, stages of meetings, and team conversations that we have found useful. In order to illustrate our general approach, we select one design meeting experience as a case study. We close with observations on how facilitation skills might be developed by design team participants. This is needed in response to an emerging requirement for effective collaborative teamwork in HCI design activities.

Modelling the user interaction methods imposed by designs

Abstract We report our experience in the use of “approximate modelling” to compare segments of two existing interactive systems. This work is part of a project investigating how the theoretical framework of Kieras and Poison for modelling complexity in human-computer interaction might be adapted and used effectively in an industrial system development environment. The process of conducting the task analysis necessary for modelling led to qualitative insights well beyond the quantitative predictions provided by applying the theory. Though the methodology can provide a useful focus for task analysis in the design process, the quality of the resulting insights was highly dependent on the set of tasks selected for analysis.

Tools and methodology for user interface development

5. Suggested topics and areas of researc h The information generated from a workshop affects th e direction of subsequent research on the topic. One of th e goals of this workshop was to suggest specific areas of use r interface and UIMS research that should be addressed b y industry and academia. In general, the predominant areas of research are th e development of tools that provide support for th e development/evaluation of user interfaces, the developmen t of higher-level languages for the definition an d implementation of user interfaces and the development o f formal taxonomies/categorizations of UIMS. Nine specifi c research topics are as follows : 1) Determine how artificial intelligence and knowledge engineering can support the definitions of th e conceptual model of the users problem or system bein g developed. In the development of a system such as a user interface, research should define a method, tool or environment to enable a knowledge engineer to presen t the end user with a description or conceptual model o f the system. Modeling is particularly important in th e development of a user interface. 2) Define and develop data structures that will provid e adequate support and effective interfaces for use r interface software tools. 3) Develop user interface design tools that can acquire an d use higher-level knowledge about user interfac e concepts for particular application domains. Thi s higher-level knowledge includes the objects, actions an d relationships in specific problem domains. Design tool s should allow the definition and reuse of primitives i n specific application domains, thus enabling user interface designers to begin defining user interfaces i n terms of abstract objects instead of lower-leve l programming language concepts. 4) Determine the characteristics of good user interfac e designs and how these characteristics can be quantifie d and measured. Tools should then be developed that will encourage user interface designers to define goo d user interfaces by making easy those decisions that result in good user interfaces and making difficult thos e decisions that result in bad user interfaces. 5) Develop prototyping tools that will facilitate feedbac k from the end users of the interface. The feedback resulting from the end users interaction with th e prototype can be used to further refine the requirement s of the user interface. 6) Develop a formal taxonomy or categorization of UIM S functionality with respect to …

Report on the workshop on analytical models

Analytical models are extremely useful and important tools for design engineering in many fields. They provide engineers with performance predictions for a specified design, and they also can assist in design improvements by identifying the sources of performance deficiencies.Currently analytical models are not commonly used for design engineering in HCI. However, university research has produced some promising results, and has led to exploratory development projects in several major corporations.