“…It, however, turns out, see [17], that the idea of factorization that has successfully led to the construction of essentially optimal multigrid solvers in [10][11][12][13][14][15][16] cannot be applied to an important class of schemes that is routinely used for compressible flow computations -the shock-capturing schemes of all types, including the original formulation of genuinely multidimensional scheme, see [7][8][9]. As these schemes possess a collection of features that cannot be either compromised or traded for in computational practice, primarily the mechanism for handling discontinuous solutions, there is a need for discretization that would rather add the capability of factoring the governing PDEs to the existing properties of the scheme.…”