Abstract
One of the main problems to be solved in design-by-features is to preserve the semantic correctness of feature-based models. Currently, feature-based parametric design (FbPD) is being used as one of the most powerful approaches for solving this problem. In this paper, a fundamental principle of this approach is introduced. Three aspects stated, are: FbPD deals with functional design primitives, it solves the automatic generation of model variations, and it offers the basis for the development of a mechanism to check the semantic correctness of feature-based models. Several concepts for the definition of semantic constraints are presented. They instigate the classification of semantic constraints in four different categories, based on the constraint evaluation-time, purpose, behaviour, and representation. Sinfonia, a system for feature-based parametric design, is presented as a testbed environment for design-by-features applications. One of its modules, the Consistency Handler, uses the constraint concepts introduced in order to preserve the semantic consistency of the models. Several examples illustrate the different types of constraints. In addition, an algorithm applied for the process of a consistent feature modification is presented.