Michael Franziskus Hönig, Domenik Eichhorn
New and Controversial Ideas Track
 | Haus der Universität, Schlösslistrasse 5, 3008 Bern, Switzerland |
 | 7 February 2024, 16:20 CET |
 | Michael Franziskus Hönig |
 | Lukas Güthing |
 | https://dl.acm.org/doi/10.1145/3634713.3634735 |
Applications for feature models, such as sampling, usually involve exploring a decision structure to systematically generate product configurations. This decision structure is often learned implicitly, using SAT solvers, or explicitly by describing it in the form of a binary decision diagram. Another structure, context-free grammars, have only been discussed for constraint-free feature models. We outline two algorithms that allow the transformation of feature models into context-free grammars and argue that, though those initial algorithms do not perform well, context-free grammars show promising potential for optimizations.