Software development is increasingly conducted in a recurrent fashion, where the same product or service is continuously being developed for the marketplace. Still, we lack detailed studies about this particular context of software development. Against this backdrop, this dissertation presents an action research study into Software Inc., a large multi-national software provider. The research addressed the challenges the company faced in managing releases and organizing software process improvement (SPI) to help recurrently develop and deliver a specific product, Secure-on-Request, to its customers and the wider marketplace. The initial problem situation was characterized by recent acquisition of additional software, complexity of service delivery, new engineering and product management teams, and low software development process maturity. Asking how release management can be organized and improved in the context of recurrent development of software, we draw on Pettigrew’s contextualist inquiry to focus on the ongoing interaction between the contents, context and process to organize and improve release cycle practices and outcomes. As a result, the dissertation offers two contributions. Practically, it contributes to the resolution of the problem situation at Software Inc. Theoretically, it introduces a new software engineering discipline, release cycle management (RCM), focused on recurrent delivery of software, including SPI as an integral part, and grounded in the specific experiences at Software Inc.

Software changes constantly and continuously. It is often beneficial to record the progressive changes made to software, so that when any problems arise, it is possible to identify the change that might have caused the problem. Also, recording these changes enables recovery of the software as it was at any point of time. A version control system is used to track modifications to software. Version control systems (VCS) display when and where a change was made. In the case of multiple developers working on the same software system, version control systems also record which developer was responsible for the change. RCS, SCCS and CVS are examples of such version control systems, and they usually have a command-line interface. The widespread use of CVS has however given rise to a host of "CVS clients", which provide a two-dimensional graphical interface to CVS. While working with a version control system in two dimensions is a definite improvement over traditional command line interfaces, it is still not sufficient to display all the necessary information in a single view. Using three dimensions to display the information from a version control system like CVS is an effective and efficient way to represent multiple attributes in a single view. There are many advantages to using a third dimension for visualizing the version history and evolution of software. A three-dimensional visualization tool has been developed to provide insights into the structure and characteristics of the history of a software system. It demonstrates the benefits of three-dimensional visualization and illustrates a framework that can be used to automatically derive information from a version control system. / Master of Information Systems