A Customised ASM Thesis for Database Transformations

Abstract

In order to establish a theoretical foundation for database transformations, we search for a universal computation model as an umbrella for queries and updates. As updates are fundamentally distinct from queries in many respects, computation models for queries cannot be simply extended to database transformations. This motivates the question whether Abstract State Machines (ASMs) can be used to characterise database transformations in general. In this paper we start examining the differences between database transformations and algorithms, which give rise to the formalisation of five postulates for database transformations. Then a variant of ASMs called Database Abstract State Machines (DB-ASMs) is developed, and we prove that DB-ASMs capture database transformations, i.e. the main result of the paper is that every database transformation stipulated by the postulates can be behaviourally simulated by a DB-ASM.