Abstract: "The major advantage of a deductive database is the ability to write queries and programs declaratively, using both facts and simple logical rules to represent knowledge. Declarativeness makes queries easier to write, and thus reduces the time taken and the programming skills needed t...

Ausführliche Beschreibung

Bibliographische Detailangaben

1. Verfasser:	Ross, Kenneth A. 1936- (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	Stanford, Calif. 1991
Schriftenreihe:	Stanford University / Computer Science Department: Report STAN-CS 1386
Schlagworte:	Deductive databases Semantik Deduktives Datenbanksystem Hochschulschrift
Zusammenfassung:	Abstract: "The major advantage of a deductive database is the ability to write queries and programs declaratively, using both facts and simple logical rules to represent knowledge. Declarativeness makes queries easier to write, and thus reduces the time taken and the programming skills needed to specify a query. Two important questions arise: What should the semantics of such a deductive database language be, and how should declarative queries be answered efficiently? We address both of these questions. The question of semantics becomes complicated when one needs some sort of negation by default. For example, the absence of a fact stating that a student is in a class should allow us to conclude that the student is not in that class Several researchers have previously looked at this problem and have tried to come up with formal definitions of an intuitively satisfying semantics. Unfortunately, they were either too weak and didn't allow sufficiently many conclusions to be drawn, or they were only well-defined for a restricted class of programs. Our first contribution is a semantics that we call the 'well-founded semantics' in which all programs can be accomodated in a satisfying manner, even when the rules are recursive through negation. This semantics uses a three-valued logic in which propositions may be true, false or undefined. We study various properties of this semantics, and show how it generalizes previous approaches We also investigate how it extends to programs with additional features, such as second order contructs. As part of the second question, we look at various ways of evaluating the well-founded semantics. We propose a top-down method that is sound and complete with respect to the well-founded semantics. We investigate optimization techniques such as magic sets and show how to apply them to a large class of programs that we call 'modularly stratified.' Our magic sets method has been implemented as part of the NAIL! system at Stanford. We also investigate other optimization techniques, such as tail-recursion elimination.
Beschreibung:	Stanford, Calif., Univ., Diss.
Beschreibung:	XI, 156 S.