Pattern-based documentation analysis for software product lines

Introducing product line engineering is an investment that has to be planned for. This investment also comprises an investment into the scoping activities. Scoping, also called product line planning can be defined as the process of deciding in which parts of an organisation's products, features and domains systematic reuse is economically useful. The basis of scoping is the domain knowledge of domain experts. In existing product line engineering approaches, the scoping information is elicited interactively from domain experts, which is an effort intensive task due to the large number of experts that are involved and the time needed for workshops and interviews. In this dissertation we present the CaVE approach (Commonality and Variability Extraction) as a solution to this problem. The approach supports scoping and the introduction of product line engineering in a development organisation by systematically eliciting scoping information from user documentation. Our main goal is expert load reduction, to improve the feasibility of product line introduction in industrial context. We could show the completeness and correctness of the approach in three industrial case studies and one experiment. With the empirical validation we could show an expert load reduction of over 40% percent when using the approach in a real product line context.