Nodal DG trace finite elements

[lindsayad]

The current DG trace finite element appears modal given the diagonal mass matrix I encountered for k = 1 on 1D facets. Are there any plans for implementing a nodal DG trace space?

[pbrubeck]

The default basis is the nodal basis (DOFs are point evaluation). The mass matrix is diagonal only on the edges of triangles and on tensor product cells because we use the GL nodes.

However, the modal basis can be requested

FunctionSpace(mesh, "DGT", degree, variant="integral")

[lindsayad]

Thanks @pbrubeck. How does firedrake determine what quadrature rules to use? You mentioned Gauss-Lobatto is used here. Is that always the default?