Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Capturing product line information from legacy user documentation

: John, I.

Käkölä, T.:
Software Product Lines. Research Issues in Engineering and Management
Berlin: Springer, 2006
ISBN: 3-540-33252-9
ISBN: 978-3-540-33252-7
Aufsatz in Buch
Fraunhofer IESE ()
product line engineering; documentation; product line; product family; commonality; variability; elicitation; user documentation

The development of a software product line is seldom a green field task. Legacy systems exist that serve as an information source or that should be integrated into a product line. The information needed is usually elicited interactively with high involvement by the domain experts of the application domain. As domain experts have a high workload and are often unavailable, relying primarily on high expert involvement is a risk for the successful introduction of a product line engineering approach into an organization. This chapter presents an approach for the extraction of requirements from user documentation, which gives guidance on how to elicit knowledge from existing user documentation and how to transform information from this documentation into product line models. This approach is called the PuLSE-Framework for product line engineering. We describe the metamodel that is the basis of the approach, the extraction patterns that are derived from the metamodel, and the process that guides the application of the patterns and the derivation of information relevant for building a product line. This information can be features of legacy products, parts of use cases that can be used for product line analysis, different kinds of requirements and, most important forproduct line engineering, commonalities and variabilities among existing products. With thehelp of this information, a product line model with the product line requirements can be builtmuch faster and the workload of the domain experts is significantly reduced. We performedan initial validation of the approach in industrial case studies and in a controlled experiment.