Abstract The authors describe a set of best practices that were developed to assist in the design of search user interfaces. Search user interfaces represent a challenging design domain because novices who have no desire to learn the mechanics of search engine architecture or algorithms often use them. These can lead to frustration and task failure when it is not addressed by the user interface. The best practices are organized into five domains: the corpus, search algorithms, user and task context, the search interface, and mobility. In each section the authors present an introduction to the design challenges related to the domain and a set of best practices for creating a user interface that facilitates effective use by a broad population of users and tasks.

A multi-leveled framework for user interface design guidelines of Web applications is presented. User interface design guidelines tend to provide information that is either too general, so that it is difficult to apply to a specific case, or too specific, so that a wide range of products is not supported. The framework presented is unique in that it provides a bridge between the two extremes. It has been dubbed the 'Bull's-Eye' due to its five layers, represented as concentric circles. The center of the Bull's-Eye is the Component layer, followed by Page Templates, Page Flows, Interface Models and Patterns, and Overarching Features and Principles. To support this approach, requirements were gathered from user interface designers, product managers, UI developers, and product developers. Also, usability testing of the guidelines occurred on several levels, from broad guideline tests to more specific product tests. The guidelines and lessons learned are intended to serve as examples for others seeking to design families of Web applications or Web sites.

This paper describes both the process and outcome of an effort to develop an information design for a community network's web site. Typically, community networks are developed by a small group of skilled computer users who are willing to volunteer their time. This often means excluding individuals who are community-oriented but not sophisticated computer users. To encourage the participation of these community members, and to do so under the very realistic conditions of zero funding for research, we applied user-centered design techniques and drew from diffusion of innovations literature to develop the web site. This combination of techniques led to targeting innovators and early adopters in the community. Low-fidelity prototypes were used in a focus group setting, and usability tests were employed to create the initial web site information design.

This experiment examined how varying implementations of two user-centered design principles affected the usability of a computer program. Drawing from human-computer interaction, as well as cognitive psychology, the following principles were implemented and tested: maximizing use of the user's expectations and stereotypes and minimizing the user's memory load. Nine interfaces were created using Hypercard' s datebook program. On each interface, the graphics and text were manipulated according to the design principle and the related cognitive psychology research. A total of 108 people participated. Of the five hypotheses, only one was supported by the data. The data suggest that for less complicated programs and tasks, strict adherence to design principles may not be required.