For second-order partial differential equations (PDEs), approximants that possess constant
and linear precision are sufficient for convergence in a Galerkin method [20].
2. The {i (x)}n
i=1 are sufficiently smooth (typically C∞) in the interior of C.
3. Convex combination