I have this question related to the two interesting specialcf for a very long time…
In particular, I want to check the specific normal direction of an edge in a triangular mesh.
For the mesh size function
h = specialcf.mesh_size,
I can plot it on a triangular mesh as a standard CoefficientFunction, and it looks like (should be) the area.
I guess, when restricting to the edges, it became the edge-length, right? But I can not verify this claim since I don’t know now to evaluate a Coefficient Function at an edge manually… Here is a related question concerning this issue: how to manually evaluate a function that is only defined on the edges? like the stabilization parameter for an HDG method, or its skeleton solution? The approach of MappedIntegrationPoints
mip = mesh(0.2,0.2)
seems not work for edge-wise evaluations…
Then, for the normal direction, the question is similar, I want to know n and n, manually, at all the mesh skeleton…