| This dissertation discusses the adaptation of object-oriented features, such as object identity, complex objects, methods and inheritance, to first-order logic in the context of deductive object-oriented databases. Specifically, it develops a translation algorithm to transform a deductive object-oriented language, e.g., the F-logic language, into a first-order logic language. Development of the translation of programs enables us to take advantage of the theories and techniques of first-order databases for query evaluation to the fullest extent.; The algorithm of translation is for the monotonic part of the deductive object-oriented language. The system uses a model theory to show that the translation is possible by either direction. When inheritance overriding is present, a model-theoretic semantics and a fix-point theory are shown to compute the intended model and correct answers. This approach also proves that the intended model is minimal and unique.; The system also uses a top-down approach which is based on SLD-resolution to evaluate the rewritten programs. Inference rules due to object-oriented-features are adapted and formulated. It is shown that the resolution is sound and complete.; The study concludes that the expressive power of first-order logic theory can be well incorporated with the modeling features of object-orientation. The semantics of inheritance in object-oriented databases can be given in terms of rule deduction. This dissertation also casts the other object-oriented features in the workings of predicate logic. The investigation, then, reveals the integration of a first-order database with object-orientation. |
