Unified Database Theory

The paper provides a common mathematical database theory that unifies relational databases, data warehouse models and object-oriented database models. It is based on long experiences reflected in human language models and database system development. This is a short version of the theory. The complete version can be requested at [email protected]. The paper discusses a state model as the base for storing data in a database. This includes specific ways of collecting states in instances and instance collections as well as ordering states in classes by means of classifications. Thus, database models can be described on four different levels according to the degree of order in the database. The paper investigates how far principles of presenting knowledge in human language can be used as basic rules for building database systems. It shows a way to formalize knowledge presentation in human language and how human language principles as ordering facts, referring to behaviour and causalities can be used for building database models. Theoretical and practical benefits of human language oriented DBMS (database management systems) are demonstrated. The paper is based on practical experiences in developing database management systems (mainly the object-oriented DBMS ODABA) with the goal of supporting human language oriented knowledge presentations.