Jørgen Schartum Dokken

This seems to be supported by other packages using VTK, such as pyvista, (ref minimal example in https://github.com/pyvista/pyvista/issues/947). Is there anything in particular missing in itkwidgets for this?

Fixed in #2323

@tiagovla Feel free to add such a contribution to the code. As far as I can tell by quickly looking at: https://raw.githubusercontent.com/Kitware/vtk-examples/gh-pages/src/Testing/Baseline/Cxx/GeometricObjects/TestLinearCellDemo.png and https://github.com/FEniCS/basix#prism The map for a first order...

I would strongly suggest that you use: `compute_global_tree` and `compute_collisions_point` to reduce the number of points shared with each process. See for instance: https://github.com/FEniCS/dolfinx/blob/d316ef77c58883f09d0e68486aa1caa24a10f6ec/python/test/unit/geometry/test_bounding_box_tree.py#L431 With such an implementation, you can...

Imagine you have two sets of meshes with 500 millions vertices in each (and assuming that we have a CG-1 space that would be the same number of dofs). If...

We would need some tests to make sure this is correct. I.e. Interpolate a known function onto a sub-set (similar to your C++ demo), and check that the interpolated function...

> > There are questions about scalability of this code. When every process communicates with every other process, it becomes a quadratic scaling, which is not good. Obviously, it can...

> Simplicity. Is there a compelling reason/need to support `int8_t` and `int64_t` in IO? > > > > Is there a current use case for double? I was currently planning...

@garth-wells we do not

Some consistency and usability has been fixed in #658 . The tests are quite useless, as they don't really test the io, and consumes quite a lot of time.