User-centered application definition: a methodology and case study - includes related articles on interviewing and storyboarding techniques

Hewlett-Packard Journal, Oct, 1989 by Lucy M. Berlin

User-Centered Application Definition: A Methodology and Case Study

GOOD PRODUCTS MEET USERS' NEEDS, present their interactions in terms of their users' model of operations, and are easy to use. They accomplish this by embodying knowledge about the user's tasks and sequencing of operations, and about the relationship of the tasks to the overall work context. For example, VCRs can be difficult to use because they have an obscure tape counter, a control that reflects the internal structure of the machine, rather than a "minutes of playing time" indicator which would reflect the user's view of the information. In contrast, a washing machine is easy to use, since the user simply indicates the type of clothes and the size of load, and is not required to understand the machine's internals or to specify the hot water flow rate, gallons of rinse water, and spin rate of the spindle. Good products present user interfaces that are consistent with the user's model of their operation rather than mapping too directly to the product's internal organization.

Software applications, and in particular their user interfaces, also need to be consistent with their users' model of operations. Dramatic advances in hardware integration and an exponential decline in costs have expanded computer use beyond the original computer science and engineering domains into many new contexts. As a result, software users can no longer be assumed to be computer experts who are comfortable with the traditionally terse user interfaces. These noncomputer specialists are too often required to translate between their professional domain and the restrictive terms of the computer. For them, commands such as find. - type f - mtime - 7 - exec grep - I'Hello' {} ; are viewed as magical incantations, not meaningful communication. (*1) They must try to model their needs into the system's internal structures and concepts (similar to manipulating tape counter) instead of communicating with the system in the terms of their domain and tasks. To provide good products for these new users, developers must actively seek out an understanding of their users' domain and tasks. It is this need for developers to do more to understand users and their needs that we call user-centered design.

In the project covered in this paper, we set an explicit goal to produce a user interface design that is focused on user tasks. We developed the user-centered design methodology described here to help us gain this crucial understanding of our users' needs, and now propose it as a more generally useful process. Using this approach, other developers may learn, for example, that seemingly similar tasks in different domains or involving different users may actually have very different, domain-specific requirements for functionality and user interface. For example, computer-aided design applications for VLSI circuits and for mechanical engineering appear similar at first glance. Both systems use hierarchical, multiple-view design representation data structures and graphical editing. However, these two domains use very different design verification methodologies, each with its own optimal user interface. Each system is most useful if the users can interact using their own domain concepts, if common tasks are easy to perform, and if the systems fit into the specific work context. To be done well, both applications require that the developers go beyond their own experience and acquire an understanding of the problem space from their users' point of view. The methodology described here may help developers acquire this understanding and provide systems and user interfaces that more closely match their users' task models.

Since we use our project's hypertext platform definition process as a case study of the application of the user-centered methodology, we briefly describe the project's background and goals in the next section. The following section defines the user-centered methodology we developed and how it relates to other work. Each succeeding section details a single step of the methodology, and uses our project as a case study of how the step was applied in practice. We end with a discussion of the project's current status and a retrospective view of the contributions and general applicability of the methodology.

Project Background and Goals

The information interfaces project is in the Human-Computer Interaction Department of the Software Technology Laboratory at Hewlett-Packard Laboratories. The Software Technology Laboratory develops environments, tools, and methodologies to improve software productivity.

Our project's charter is to explore information management issues and technologies. We hypothesized that hypertext is a useful enabling technology in many applications that access complex information. Our approach was to demonstrate the power of hypertext by defining and prototyping a hypertext-based software platform. A software platform is a common toolkit of concepts and capabilities for use by developers of many applications. The hypertext capability makes available nonlinear text with connections (called links) among related items.