Interactive techniques to support the configuration of complex feature models.

Whenever a software engineer derives a product from a product line, he has to resolve variability by making con guration decisions. This con guration process can become rather complex because of depen- dencies within the variability model and knock-on e ects and dependencies in other related artefacts. Because of the limited cognitive capacity of the human engineer, this complexity limits the ability of handling product lines with large con guration spaces. To address this problem we focus on techniques that support the interactive con guration of larger feature models, including (1) visual interaction with a formal reason- ing engine, (2) visual representation of multiple interrelated hierarchies, (3) indicators for con guration progress and (4) ltering of visible nodes. The concepts are demonstrated within S2T2 Con gurator, an interactive feature con guration tool. The techniques are discussed and evaluated with feature models, however, we believe they can be generalised to other models that describe con guration choices, e.g., variability models and decision models.