Validity maintenance of semantic feature models

One of the most powerful characteristics of feature-basedmodeling is the ability to associate functional and engi-neering information to shape information in a productmodel. Current feature modeling systems embody thisparadigm in their graphical user interfaces, providingthe user with “engineering rich” dialogs aimed at thecreation of feature instances. Most systems, however, failto consistently maintain the meaning of the featuresthroughout the modeling process. For example, a model-ing operation on one feature may affect the semantics ofother features without the user being notified by the sys-tem, let alone assisted in overcoming the situation.Semantic feature modeling is a declarative modelingapproach that not only provides a well-defined specifica-tion of feature semantics, but also effectively maintainsthis semantics during the modeling process, for all fea-ture instances in the model. This paper describes thevalidity maintenance mechanisms of the semantic fea-ture modeling approach. These include (i) detecting eachinvalid situation, (ii) reporting it to the user, with appro-priate explanation on its causes and effects, and (iii) pro-viding the user with a convenient choice of reactionhints, aimed at recovering validity in the model. An ex-ample modeling session is also given, illustrating whichhigh-level user assistance is provided under this ap-proach.