Current feature-based modeling systems fail to adequately maintain feature semantics. This is partly due to inappropriate specification of validity conditions in feature classes, but mainly due to a lack of effective validity maintenance mechanisms throughout the modeling process. A critical aspect in this is feature interaction management. This paper presents a new approach to the detection of feature interactions, which uses semantic and interaction constraints in feature class specification. Validity maintenance is automatically performed after each modeling operation by checking these constraints, thus being able to detect a variety of interaction types. Such interactions are then analyzed, and their causes identified and reported to the user.

This content is only available via PDF.
You do not currently have access to this content.