Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Goal-oriented quantitative software project control

: Heidrich, Jens


Stuttgart: Fraunhofer IRB Verlag, 2009, XV, 345 S.
Zugl.: Kaiserslautern, TU, Diss., 2008
PhD Theses in Experimental Software Engineering, 24
ISBN: 978-3-8167-7932-2
Fraunhofer IESE ()
project management; project control; measurement; goal question metric approach; quality improvement paradigm; management decision making; management & management techniques

Detecting and reacting to critical project states in order to achieve planned goals is one key issue in performing a software development project successfully and staying competitive as a software development company. Software Project Control Centers support the management and controlling of software and system development projects and provide a single point of project control, which addresses distributed software development in particular. One key element of such control centers is the analysis, interpretation, and visualization of measurement data used to support various stakeholders in controlling a software development project.
Currently, many companies are developing their own control centers (typically simple dashboards). However, support for identifying suitable indicators and integrating appropriate custom-tailored control mechanisms into control centers is widely missing. Usually, these control centers provide a fixed set of predefined mechanisms and measures for project control that cannot be sufficiently customized to project and orga¬nizational specifics.
This thesis will introduce the Specula project control approach for goal-oriented composition of project control centers. The focus is on how to describe and combine control components and how to select the right components based on explicitly defined measurement goals. It provides (a) a conceptual model for formally specifying relationships between applied control components (data collection procedures, purpose-oriented control techniques, and role-dependent visualization mechanisms), (b) a logical reference architecture for implementing control centers, (c) a repository of standard control components, (d) a methodology for selecting and adapting control components based on explicitly defined measurement goals with respect to project control, and (e) a prototype tool supporting the overall approach.
The approach was evaluated in industrial case studies including 9 projects from different domains with different existing project control mechanisms in place. The validation demonstrated, for instance, that 80% of plan deviations and project risks were found earlier than when using a traditional approach to project control. Moreover, 20% of the deviations and risks found by Specula had not been detected at all by the traditional approach. Furthermore, the general usability and ease of use of the Specula approach was illustrated.