Research into Database System Theory

9223405乌尔曼对数据库系统理论的研究这是为期三年的连续奖项的第一年。高级代数或逻辑查询语言已被证明在现代数据库系统中非常有用，因此需要扩展它们的范围。这些类型的语言允许清晰、快速和准确地编写数据库程序。然而，它们使数据库系统更难找到有效的方法来回答查询或实现约束。逻辑表达的约束是分布式数据库系统的一个重要特征，特别是在用于独立数据库的协作或集成时。在连接的数据库更改时检查约束的方法尤其重要；目标是尽可能本地和高效地检查约束冲突。逻辑语言经常涉及规则中的否定，或非单调推理，其中在一个层面上提供更多的信息允许一个人在另一个层面上推断较少。人们正在寻找有效和普遍的方法来找到对带有否定的逻辑的直观正确的解释。最后，使用代数或逻辑作为查询语言需要优化涉及多个关系联接的表达式。寻找良好的加入秩序的有效方法正在研究中。如果成功，大型数据库系统将变得更容易使用，并将能够执行诸如分布式约束维护之类的操作，而这些操作是它们今天无法以足够的效率完成的。***

9223405 Ullman Research into Database System Theory This is the first year of a three-year continuing award. High- level algebraic or logical query languages have been shown to be extremely useful in modern database systems, and there is a need to extend their scope. These sorts of languages allow database programs to be written clearly, quickly, and accurately. However, they make it harder for the database system to find an efficient method to answer queries or to implement constraints. Logically expressed constraints are an important feature of distributed database systems, especially when used for collaboration or integration of separate databases. Methods for checking constraints as the connected databases change is especially important; goals are to check for constraint violations as locally and efficiently as possible. Logical languages frequently involve negation in rules, or nonmonotonic reasoning, where providing more information at one level allows one to infer less at another level. Efficient and general ways to find intuitively correct interpretations of logic with negation are being sought. Finally, the use of algebra or logic as a query language requires the optimization of expressions involving the join of many relations. Efficient ways to find good orders of joining are being studied. If successful, large-scale database systems will become easier to use and will be capable of performing operations such as distributed constraint maintenance that they are not able to do with adequate efficiency today. ***