Detecting physical and logical dependencies to support change impact
|Autoren|| Günter Pirklbauer|
|Titel||Detecting physical and logical dependencies to support change impact|
|Institution||Software Competence Center Hagenberg GmbH|
The costs for enhancing and maintaining software systems are 75% of the total development costs. It is therefore important to provide appropriate methods, techniques and tools to support the maintenance phase of the software life cycle. One major maintenance task is the analysis and validation of change impacts. Existing approaches address change impact analysis, but using them in practice raises specific problems. Tools for change impact analysis must be able to deal with inconsistent requirements- and designmodels, and with large legacy systems and systems, which are distributed across data processing centers. We have developed an approach and an evaluation framework to overcome these problems. The proposed approach combines methods of dynamic dependency analysis and change coupling analysis to detect physical and logical dependencies. The implementation of the approach - a framework consisting of methods, techniques and tools - will support both the management and developers. The goal is to detect low-level artefacts and dependencies based on only upto- date and system-conform data, including logfiles, the service repository, the versioning system database and the change management system database. With the assistance of a data warehouse, the framework will enable dynamic querying and reporting.