An interaction paradigm for impact analysis

Abstract

The Aerospace industry is concerned with huge software projects. Software development is an evolving process resulting in larger and larger software systems. As systems grow in size, they become more complex and hence harder to maintain. Thus it appears that the maintenance of software systems is the most expensive part of the software life-cycle, often consuming 50-90% of a project total budget. Yet while there has been much research carried out on the problems of program and system development very little work has been done on the problem of maintaining developed programs. Thus it will be essential to improve the software maintenance process and the environment for maintenance. Historically, the term Software Maintenance has been applied to the process of modifying a software program after it has been delivered and during its life time. The high cost of software during its life cycle can be attributed largely to software maintenance activities, and a major part of these activities is to deal with the modifications of the software. These modifications may involve changes at any level of abstraction of a software system (i.e design, specification, code,...). Software Maintenance has to deal with modifications which can have severe Ripple Effects at other points in the software system. Impact Analysis addresses the problem and attempts to localize these Ripple Effects. In this thesis the Software Maintenance process and more specifically the Impact Analysis process is examined. The different parts of the implementation for the Impact Analysis System are explained. The main results of the thesis are the dependencies generation and the graph tool used to visualize these dependencies as well as the impacts on general dependency graph for impact analysis purpose.