No abstract

The Human-Computer interface is crucial to good design support tools. It has to be non-interruptive and non-distracting, yet allow the architect to interact with the computer software.The physical reality of the interface, such as the shape and manipulability of devices like the mouse, keyboard, joystick, or dataglove, has to be mapped on actions and commands in the software. Already the current user interfaces are felt to be inadequate for a good support of design, and the functionality of design tools is growing, requiring even more and new physical interface devices. In this paper, we present research on new tangible interfaces for architectural design support. In particular, we focus on the research methodological question how to investigate such devices.The research strategy is introduced and discussed, after which concrete implementations of this strategy are shown. Based on this work, we conclude that the combination of interface and the context of its use in terms of design method and user needs form crucial aspects for such research and cannot be considered separately. 134

Bayesian belief networks offer an alternative to conventional estimation methods in estimating user preference or utility functions. Because parameter estimates are updated sequentially, this approach seems very promising in user-centred design and data collection systems. The application of such networks however poses several questions, related to speed of learning, sample heterogeneity and discretionalisation of the parameter space. This paper reports the results of a series of numerical simulations which were conducted to gain more insights into these operational decisions.

The applicability of conjoint analysis to measure housing preferences has been intensively debated in housing research. How to improve respondent understanding of the profiles and whether visual or verbal representation should be preferred are relevant issues in this context. This article reports the results of a study that was designed to assess the effects of pre-experimental training, using virtual reality, on various forms of validity. Results suggest that the use of virtual reality helping subjects to learn about the nature of the attributes that are varied in the conjoint experiment significantly improves the validity of the conjoint analysis.

No abstract