摘要： 从证明论(Proof Theory)的观点看,演绎数据库构成了一种特殊的一阶理论,并被分为确定(Definite)和不确定(Indefinite)两种形式。很容易用PROLOG表示和处理确定型演绎数据库,因为组成它的事实和推演规则(Deductive Rules)为一组Horn子句。非Horn子句(non-Horn Clauses)的存在使不确定演绎数据库向PROLOG数据库的映射复杂化。实现这种映射面临的主要问题为:1).如何用PROLOG表示和处理non-Horn子句;2).如何克服闭域假设(Closed World Assumption)下的不一致性;3).空值(Null Values处理;4).如何得到不确定解答(Indefinite Answers)。本文分别讨论了这些问题,并提出了解决方法。

Abstract: From a proof-theoretic viewpoint, a deductive database forms a special first-order theory. There are two different types of deductive databases, namely definite and indefinite [1]. The former can be easily expressed and processed in PROLOG. In fact, a definite deductive database consists of nothing else than a set of Horn clauses. However, the addition of non-Horn clauses makes it difficult to map an indefinite deductive database into a PROLOG database. The main problems faced are:l) how to express and process non-Horn clauses in PROLOG; 2) how to deal with the inconsistency under the closed world assumption; 3) how to treat null values; and 4) how to extract indefinite answers. All these problems are discussed, and some simple ways to solve them are exploited in this paper.