Taxonomic reasoning in LOGIDATA/sup +/
Domenico Beneventano, Sonia Bergamaschi, C. Sartori · 2002
The applicability of the inference technique named taxonomic reasoning to LOGIDATA/sup +/ schemas is demonstrated. The LOGIDATA/sup +/ model is enhanced with defied classes and set cardinalities. In particular, the problem of determining subsumption between classes with cyclic descriptions is addressed. Since recursive definitions are allowed, it is necessary to choose a strategy for determining fixpoints. The subsumption algorithm presented uses greatest fixpoint semantics, and this choice is compared to some alternative approaches, such as least fixpoint and descriptive.>