“…In practice, the different choices of ansatz and test functions for MOT schemes may lead to instabilities. Proper Galerkin methods are not only provably stable, but they have also attracted interest from at least three different perspectives: Rigorous a posteriori error estimates give rise to efficient adaptive mesh refinement procedures [11,15,16,17,20,25]; non-polynomial basis functions and efficient assembly of the algebraic system [21,25,24]; formulations based on the physical energy [2,3,4]. In this context, efficient preconditioners have been of current interest.…”