Thanks for the fast response and I see what you mean. But isn’t this in big contrast to the same problem in 2D?
If I want to solve let’s say a Poisson-like equation in 2D and I want to modify the parts of the boundary where to apply Dirichlet- and where to apply von Neumann-data, then in the most simple case, one only has to update the ‘def boundary_D(x, on_boundary)…’-part of the code. The bilinear form itself as well as the surface part remains unchanged.
In 1D, I not only have to change the ‘def boundary’-section but additionally also the “surface-part”,which of course consists of only two single points in this case.
I thought, fenics would handle this issue somehow automatically, since the way you write down the linear and bilinear forms are very reminiscent of the multi-dimensional formulation of the PDEs of course and there nothing changes…
Anyway… thanks again for this clarification!