Configuration management using SySL

Abstract

Software configuration management is concerned primarily with the consistent labelling and tracking of project information and managing change to that information [l]. Its objective is to try and control the changes that are made to the software in such a way as to preserve the integrity of the system and provide a basis on which to measure quality, both of the system and the development process. Initially, software configuration management systems were aimed at the results of the software life-cycle itself ( i.e. software components ) without trying to manage other outputs from the various stages in the process. Systems such as Make [2] and SCCS [3] are indicative of such an approach. However recent research has recognised that software configuration management should be applied throughout the software process. Therefore information such as design diagrams, requirements documents, test data, etc, as well as code, should be under the control of the configuration management system. System modelling lies at the heart of software configuration management. The system model captures the state of the system, identifies the parts making up the system and specifies how to put these components together. To provide features such as the identification of baselines, the system model must capture the process by which the system is constructed, the identifiable pieces of information which emerge during the development process and their relationships. There are many dimensions to system modelling identified in [4].