nalgebra f64 interop

Right now flight uses `f32`s everywhere internally. This makes a lot of sense because the GPU only uses 32-bit floats. However, many applications that incorporate physics would rather use 64-bit transformations. Ideally, nalgebra should have some kind of `cast` function that users can use to convert ... does it?

Start by checking how flight represents f32 transformations and whether nalgebra already provides a cast operation for converting between f32 and f64 types. The issue does not name files or tests, and completion criteria would need to establish the intended interop design and supported conversion behavior.

Written by the indexing model from the issue text.