> Their code has the implicit assumption that all elliptic problems have homogeneous Dirichlet boundary conditions!