Request for Opinions on a Special Operator
Hi,
I have a question on a special kind of operator and how complicated you would judge it (or something like it) to be added to HOG.
In HyTeG's TerraNeo module there is a [RadialProfile](https://hyteg.pages.i10git.cs.fau.de/hyteg/structterraneo_1_1RadialProfile.html) class which computes information for an FE function (currently only for the P1 and P2 families) over each radial shell of the micro mesh. One such information is the **mean**. Currently this is actually only the mean value of the degrees of freedom, so not really FEMish.
What I'd actually wanted for some FE function $u$ is the actual mean. On a shell $S_r$ with radius $r$ this would be given by
```math
u_{\text{mean},r} = \frac{1}{|S_r|} \int_{S_r} u(x) \, dx = \sum_{k=1}^N \nu_k \int_{S_r} \phi_k(x) \,dx
```
when $u$ has $N$ degrees of freedom $\nu_k$ for the basis functions $\phi_k$. I know we can already do stuff on manifolds. My need would be for an "operator" $\mathcal{M}$ with
```math
\mathcal{M}: \mathcal{P} \rightarrow \mathbb{R}^{L}\enspace,\quad u \mapsto M\cdot\left(\begin{array}{c} \nu_1 \\ \nu_2 \\ \vdots \\ \nu_N \end{array}\right)
```
where $\mathcal{P}$ is an FE function space, $L$ is the number of micro shells and the matrix $M\in\mathbb{R}^{L\times N}$ is given by
```math
(M)_{ij} = \frac{1}{4\pi r_i^2} \int_{S_{r_i}} \phi_j(x) \, dx
```
What's your take on the possibility of and the difficulties in getting this to work?
issue