A Study on Development of a Deductive Object-Oriented Database and Its Application to Genome Analysis

A deductive object-oriented database (DOOD) is the integration of a deduc-tive database (DDB) and an object-oriented database (OODB). DDB and OODB are next generation databases proposed for overcoming the shortcomings of relational databases (RDB). The DDB is an extension of the RDB. It is based on the rst order predicate logic, and provides a declarative query (and programming) language. An advantage of the DDB is that the recursive query processing can be easily expressed. However, it is associated with the following problems. Many methods have been proposed for recursive and negative query processing. However, there are few systems implementing these methods and almost no practical applications. Therefore, it is not clear whether the methods are applicable to and useful for practical applications. The DDB can handle nested relations by using functions. However, nested relations are not sucient for representing the data in practical applications. The OODB provides a data model that can represent a variety of data used in practical applications. Recently, many OODB management systems have been developed and used in practical elds, such as computer-aided design, manufacturing and engineering. However, the following weaknesses have been noted. A lack of declarative query languages: Application programs should be written in procedural languages, such as C++ and Smalltalk. Diculty in representing and managing incomplete knowledge: Users must write application programs in procedural languages to handle incomplete knowledge, and it is dicult to manage them in the database. The DOOD was proposed to integrate the DDB and OODB. It was intended to possess the advantages of both the DDB and the OODB; that is, declarativeness and an object-oriented data model. Current problems with the DOOD are that: 1 there is no consensus for a framework for integrating these two databases ; it is not clear what kind of techniques in the DDB are necessary for the integration. It was with the above issues in mind that this project was initiated. First, we developed a prototype of a query evaluator of a deductive database and applied it to the following practical problems. a secondary structure search in RNA sequences of human genome, and conguration detection problems in CAD systems of petrochemical plants. The genome of an organism consists of its set of chromosomes, which includes all of its genetic information. Genome analyses require at least the following three functions: 1) management of a huge volume of data produced by experiments, 2) representation of …

演绎面向对象数据库（DOOD）是演绎数据库（DDB）和面向对象数据库（OODB）的集成。DDB和OODB是为克服关系数据库（RDB）的缺点而提出的下一代数据库。DDB是RDB的扩展。它基于一阶谓词逻辑，并提供一种声明式查询（和编程语言）。DDB的一个优点是递归查询处理可以很容易地表达。然而，它存在以下问题。已经提出了许多用于递归和否定查询处理的方法。然而，很少有系统实现这些方法，并且几乎没有实际应用。因此，不清楚这些方法是否适用于实际应用以及对实际应用是否有用。DDB可以通过使用函数处理嵌套关系。然而，嵌套关系不足以表示实际应用中的数据。OODB提供了一种可以表示实际应用中使用的各种数据的数据模型。最近，许多OODB管理系统已经被开发出来并用于实际领域，如计算机辅助设计、制造和工程。然而，已经注意到以下弱点。缺乏声明式查询语言：应用程序应该用过程性语言编写，如C++和Smalltalk。在表示和管理不完全知识方面存在困难：用户必须用过程性语言编写应用程序来处理不完全知识，并且在数据库中管理它们很困难。提出DOOD是为了集成DDB和OODB。它旨在拥有DDB和OODB两者的优点，即声明性和面向对象的数据模型。DOOD当前的问题是：1）对于集成这两种数据库的框架没有达成共识；不清楚DDB中的哪些技术对于集成是必要的。考虑到上述问题，启动了这个项目。首先，我们开发了一个演绎数据库查询求值器的原型，并将其应用于以下实际问题：人类基因组RNA序列中的二级结构搜索，以及石化厂CAD系统中的配置检测问题。一个生物体的基因组由其染色体组组成，其中包括其所有的遗传信息。基因组分析至少需要以下三个功能：1）管理实验产生的大量数据，2）表示……