Advice-giving systems for selecting interaction objects

Abstract

In the development of interactive systems, the selection of interaction objects refers to the process of identifying usable widgets for each piece of data forming the system according to various parameters. This process can become repetitive and tedious for data intensive systems where data is abundant. To overcome this shortcoming and to provide assistance to designers who are responsible for conducting this process, two types of advice-giving systems for selecting interaction objects have been developed. SELECTVISION consists of a system where parameters are input manually and the possible selected interaction objects are displayed interactively. SEGUIA consists of a knowledge based system where the selection process can be conducted either automatically after providing a file of parameters or in a mixed-initiative approach where both the system and the designer are interacting to conduct the process. Both systems share a common internal representation of usability guidelines for selecting interaction objects expressed as a selection tree of rules. This structure has been shown to be very observable and easy to process, while preserving some maintainability.