Exposing the Myth: Object-Relational Impedance Mismatch is a Wicked Problem

Abstract—Addressing a problem of software integration is a fact of life for those involved in software development. The popularity of both object and relational technologies means that they will inevitably be used together. However, the combination of these two technologies introduces problems. These problems are referred to collectively as the object-relational impedance mismatch. A mismatch is addressed using one or more mapping strategies, typically embodied in a pattern. A strategy is concerned with correspondence between the schema of a relational database and an object-oriented program. Such strategies are employed in mapping tools such as Hibernate and TopLink, and reinforce the received wisdom that the problem of object-relational impedance mismatch has been solved. In this paper, we observe that it is not clear whether each strategy, as one possible solution, addresses the cause or a symptom of a mismatch. We argue that the problem is not tame and easily resolved; rather it is complex and wicked. We introduce a catalogue of problem themes that demonstrate the complex nature of the problem and provide a way both to talk about the problem and to understand its complexity. In the future, as software systems become more complex and more connected, it will be important to learn from past endeavours. Our catalogue of problem themes represents a shift, in thinking about the problem of object-relational impedance mismatch, from issues of implementation towards an analysis of cause and effect. Such a shift has implications for those involved in the design of current and future software architectures. Because we have questioned the received wisdom, we are now in a position to work toward an appropriate solution to the problem of object-relational impedance mismatch. Keywords—object-relational; impedance mismatch; wicked problem; problem theme I