Lehrgebiet Informationssysteme

FB Informatik

LG IS

AG DBIS

(C) AG DBIS

Grand Tour of Concepts for Object-Orientation from a Database Point of View

Nelson M. Mattos

IBM Database Technology Institute
Santa Teresa Laboratory, 555 Bailey Ave.
San Jose, CA, 95150 USA
e-mail: mattos@stlvm14.vnet.ibm.com

Klaus Meyer-Wegener

University of Erlangen-Nuremberg
IMMD VI, Martensstrasse 3
8520 Erlangen
e-mail: Klaus.Meyer-Wegener@informatik.uni-erlangen.de

Bernhard Mitschang

University of Kaiserslautern
Department of Computer Science
P.O. Box 30 49
6750 Kaiserslautern
Federal Republic of Germany
e-mail: mitsch@informatik.uni-kl.de

Full paper (postscript version, compressed by gzip or PDF version )

Abstract

Over the last few years, object-orientation has gained more and more importance within several disciplines of computer science (e.g., programming languages, knowledge engineering, and database systems). Numerous papers have defined one or another of its underlying concepts (sometimes in quite different ways), and some systems have been developed following those heterogeneous definitions. Nevertheless, papers investigating the dependencies and degrees of freedom of these concepts are rarely found. For this reason, the goal of this paper is not to add yet another definition of object-oriented concepts, but to identify existing relationships among these basic concepts that allow one to cover and classify various conceivable combinations of these conceptual building blocks. Dependencies, orthogonalities, and relations among concepts like object identity, encapsulation, classification, generalization, inheritance, etc. are revealed, showing numerous ways to compose different shades of object-orientation. This leads to alternatives encountered when constructing object-oriented systems, which are illustrated by classifying some well-known systems and prototypes from different areas. However, it is not our purpose to analyze the relative importance of these concepts. Instead, we investigate the concepts from a neutral point of view, presenting (but not evaluating) several degrees of object-orientation.

in: Data & Knowledge Engineering 9 (1992/93), pp. 321-352, North Holland