Hi again, Damaggio!
A little background:
I am working on a rudimentary game-engine that I've coded almost entirely from scratch as a learning project and to support a unique game idea. I have written my own model loaders, graphics engine and shaders. I am currently using the Bullet physics engine for collision detection. I have been working on this on my spare time for a couple years now and things are going very well.
Rather than building a full blown level editor, I have been using certain naming and structure conventions in the FBX scene I build so the model loader knows what in game objects they represent. For example, if I have a mesh node called "pBox1_dynamic" my model loader understands that this is a dynamic box object, and it creates both graphics and collision information automatically.
Current problem:
Right now, I am trying to modify my loader so that I can load height-maps into my game. The current structure I have planned for my Maya scene is:
1. Create 1 parent node named "heightMap_(some name)"
2. Create 1 child node called "map" - this will store the heightmap texture in its material.
3. Create a child mesh node or node(s) that describe the terrain itself beneath the parent node.
What my loader will do:
1. Search for all nodes who's name begins with "heightMap"
2. search its children for a "map" node. This will contain the heightmap texture. This will be passed to the Bullet physics engine to create the collision heightmap.
3. use the other child mesh node(s) to draw the terrain itself, graphically.
This may sound like a lot of work but I only anticipate having one large heightmap per level. My approach for modeling the terrain will generally be:
1.start with an existing heightmap (probably one I created in photoshop) use the "attribute map" section of the sculpt
geometry tool to apply it to a polygon plane.
2. make some minor modifications to the terrain as I edit my-scene, probably with the sculpting tool.
3. generate a new heightmap for the new terrain (using the methods we discussed in my other thread)
So up to now all that I have working, the problem is the size of the terrain.
My engine contains optimizations that avoid rendering meshes who's bounding boxes are outside of view for the current camera. For this reason, I would like to cut my final terrain into multiple mesh objects so that my game engine can avoid rendering large sections of the terrain that are not in view. See the picture below. Breaking up the mesh in this way can be done as a final step/optimization before exporting to FBX so I don't necessarily need to be able to sculpt the mesh after its broken up, but I do need to be able to partition it without modifying the texture mapping. I wouldn't want, for example, 4 copies of the blue circles to appear on each of the new meshes. If done right, the new 4 meshes would not be merely cut polygons or shells but distinct mesh nodes entirely.

Hope that makese sense, I know my purpose may be a bit obscure but I appreciate whatever ideas you have. If worse comes to worst I may need to just program my loader to break it up myself on load.