Jack Lovell

This is a general problem with Python's multiprocessing module and Numpy: we see it at Culham with codes completely unrelated to Cherab and Raysect too, when codes using multiprocessing to...

A quick scan through the code base suggests this is only an issue in the caching functions. Although np.linalg.solve appears in the interpolators too, it's only used when building the...

> A quick scan through the code base suggests this is only an issue in the caching functions. Although np.linalg.solve appears in the interpolators too, it's only used when building...

It sounds like the new interpolators would fix both this and #50 then. Is this the work going on in the [raysect/source/feature/interpolation](https://github.com/raysect/source/tree/feature/interpolation) branch? Depending on how much outstanding work there...

I don't think we should pollute the demos with additional environment variables: it adds noise and makes them more confusing for new users. The demos produce the correct output at...

I've created #363 to re-write the caching objects using the new utilities provided by raysect for the interpolation. Once that's approved and we've deprecated the original Cherab interpolators we can...

After discussing this earlier, it was agreed that we should remove the CSG aperture capability from the BolometerSlit object, as it's too error prone (this being just one example of...

I don't see any harm in doing this. You should ensure that the documentation is updated too, to include these properties.

You can make something like a `vector2d.Discrete2DMesh` already using `raysect.core.math.function.vector3d.FloatToVector3DFunction2D`. Just use separate `float.Discrete2DMesh` functions for each of the x, y and z components of the vector fields. OK, performance...

Maybe I'm missing something here, but why do we need `ModelManager`s at all? What's wrong with a standard Python list? The existing `ModelManager` implementations seem to just be wrappers around...