Intentions of Operations – Characterization and Preservation

Abstract

Frequent changes of software requirements imply changes of the underlying database, like database schema, integrity constraints, as well as database transactions and programs. Tools like ERWin, DBMain and Silverrun help developers in applying these changes. Yet, the automatic derivation might pose a problem: Since the developer is not aware of the details of the derivation applications, the resulting programs might include contradictory actions. That is, intentions of programs might be reversed by the automatic derivation, resulting a different behavior than expected by the developer.

In this paper, a compile-time algorithm that achieves preservation of intentions is suggested. The algorithm revises a composite program into a program without contradictory actions. It is based on a fine analysis of effects, that is sensitive to computation paths. The output program is expressive and efficient since it interleaves run-time sensitive analysis of already reduced effects within the input program. The compile-time reduction of effects accounts for the efficiency; the run-time sensitivity of effects accounts for the expressiveness. The novelty of the proposed approach is in combining static and dynamic analysis in a way that run-time overhead is minimized without sacrificing the expressivity of the resulting program.