On recursive axioms in deductive databases (Q1062477)

One of the basic problems of deductive relational databases is that of termination, since recursive axioms may lead to infinite derivation paths. Several methods have been proposed how to handle this problem [cf. e.g. \textit{C. L. Chang} [On evaluation queries containing derived relations in a relational data base, Advances in Data Base Theory, Vol. 1, 235-260 (1981; Zbl 0508.68061)], \textit{S. C. Shapiro} and \textit{D. P. McKay} [Inference with recursive rules, Proc. Nat. Conf. AI, Stanford Univ., Calif., 151-153 (1980)], \textit{S. A. Nagvi} and \textit{L. J. Henschen} [Performing inferences over recursive data bases, Proc. Nat. Conf. AI, Stanford Univ., Calif., 263-265 (1980)], and \textit{R. Reiter} [On structuring a first order data base, Proc. Canadian Soc. Comput. Stud. Intelligence, Second Nat. Conf. (ed. R. Perraut), Toronto (1978)]. The present paper generalizes Chang's solution to cover a larger class of cases. Moreover, a new method is suggested based on Reiter's approach - this method relies upon an adequate separation between generation and derivation rules.