I'm just curious if there as any reason why artists are so resistant to split their 3D models into multiple fbx files. I just find the code design much cleaner if 3D models are split up into multiple files, rather than having to dig through bone hierarchies to choose different parts of a model the code needs to control. Something similar is the 2D project layout with characters broken into different sprites, and being able to grab those by name. (I get that the artist can name bones to make the search nicer, but a search still needs to be done instead of giving a path) So, I get that it has to do with animating movement, but why can't that be split into multiple files? (For a maya workflow at least)
Splitting meshes would work as you intend only if every polygon in the mesh is influenced by no more than 1 bone.
If you want the model to bend smoothly polygons need to be influenced by more than one bone so your different parts will have more than one bone and require a search just the same.
Here's an example of bone-weighting with just 5 of the bones selected (in colour)
If we look closer at the pelvic area we can see a triangle affected by 4 bones at once:
Isolated we can see the blend of 4 bones represented by 4 colours:
This means it's not possible to break down the mesh strickly by bones attached.
It would also cause a performance issue in most cases as each batch of triangles would need to be drawn in a separate draw call in order to change the set of bones affecting the part currently being rendered.
Having the entire mesh as one object allows the GPU to deform (animate) and draw the whole mesh all at once.