Software is subjected to continuous change caused by new or revised requirements, or in response to detected bugs. Performing these changes becomes more and more severe as size and complexity of software increases. Failing to understand all impacts of a change can have unintended, or even fatal effects. Software change impact analysis is a family of techniques assisting in estimating consequences of a change, and is an important research topic. Nevertheless, until now researchers have not applied and evaluated those techniques in industry. This Master's thesis firstly reports on using impact analysis techniques during the development and maintenance of a large, commercial software called Exact Online. In order to support the software and quality engineers of Exact Online in performing impact analysis tasks, an impact analysis technique suitable for this large software system was introduced and implemented in a tool. Then, an evaluation was performed which revealed that the proposed technique was beneficial for the software engineers for predicting change impacts. Moreover, it successfully assisted quality engineers during regression testing.
