Trial Implementation Conceptual Schema Modelling Facility

本研究的成果由以下三个主题组成：(1)概念模式建模工具的试行（单线程环境）；(2)概念模式建模工具的试行（多线程环境）；(3)概念数据模型标准规范的研究(1)是本研究的核心。我们已经实现了一个数据建模工具，它不一定会出现有限的元数据结构。所使用的编程语言是program，所获得的结果发表在研讨会上或作为一篇论文发表在一本书中。(2)利用面向对象的编程语言实现了类似的建模功能，但没有得到预期的结果。我们尝试用Macintosh实现，但是面向对象的平台SOM不稳定，遇到了很多麻烦。一开始，我们尝试了面向对象的数据库Neo Access，但是这个包很难使用，我们不得不在项目中途更换数据库工具。我们的实现是一种需要充分利用多线程操作系统组件的DBMS，我们不能接受原因不明的错误。(3)针对这个研究项目，我们提出了与国际标准化组织类似的规范，但我们无法说服代表许多国家的专家。我们发现，即使明确说明了规范，也不足以说服许多国家机构。重要的是，所提出的规范必须具有某些在世界上被一致承认的理论或客观性质（例如数学）。因此，我们研究了用谓词演算处理传统数据模型的可能性，并发现数据模型可以用谓词演算公式表示。在逻辑中，公式的完备性或不完备性是众所周知的，我们可以更容易地说服专家。

The results of this research are composed of the following three topics :(1) Trial Implementation of Conceptual Schema Modelling Facility (Single thread environment)(2) Trial Implementation of Conceptual Schema Modelling Facility (multi-thread environment)(3) Investigation of the standard specification of Conceptual Data Model(1) is the kernel of this research. We have implemented a data modelling facility where not necessarily finite meta data structure appears. The programming language used was Prograph and the obtained results are published in a symposium or as a paper in a book.(2) was to implement the similar modelling facility using object-oriented programming language, but we could not get the expectedresults. We tried the implementation using Macintosh, but the object-oriented platform SOM was not stable and we met a lot of troubles.At first, we tried object-oriented database Neo Access, but this package was very difficult to use and we have to change database tool in the middle of our project. Our implementation is a kind of DBMS where full uses of multi-threaded operationg system component were required and we could not take bugs whose causes were unknown.(3) Parrel to this research project, we proposed similar specification to the international standardization group, but we could not persuade experts who represented many nations. We have found that even if a specification is clearly stated, that is not enough to persuade many national bodies.It is important that the proposed specification must has some theoretical or objective properties (e.g.mathematics) that is unanimously admitted in the world. Therefore, we investigated the possibility to treat conventional data model by predicate calculus, and have found that a data model can be represented by a formula of predicate calculus. In logic, completeness or imcompleteness of the formula is well-known and we can persuade experts more easily.