Seems these months concentrates on playing around with the Ogrekit (Ogre+Bullet+blender), Perfect combination actually!! Obvious advantages are:
Pros:
1. no need to worry rendering issues since everything take care by OGRE, powerful!
2. no need to take care physics stuffs too, by powerful bullet!
3. tempting 3D scene management tools from Blender side.
Although this dreamteam-like dev environment, but the fatal reason for it won't success is the capatibility to the real Blender, most likely the BGE (blender game engine) won't be supported well. So, the strong side for accessing the BGE, plus Python is completely vanished!!!! You can see the emerging for BGE and tons of millions examples to build games relying on BGE currently on the street.
So..............the point is........1.) we need to breakthrough the bottle neck and need to figuring out what/what's not supported by the blender engine on Gamekit...
Otherwise, we need to move BULLET and OGRE out and use another way round for the Physics and Rendering Engine.
then,......I wonder TORCS!!
Well, it's not easy to develop 3D scene and car mesh. OH!! May not, we can still use BLENDER to build car, road meshes and import to TORCS.
And we also need to adapt the rendering engine (OpenGL??) on Torcs to OpenGL-ES on Android.....Control from Glut/Freegluts to Android?? Don't konw.
But the point is the flexiblity will be increased for that.
Anyway, this is plan B!
PS: Athough Gamekit is a good game engine, but the documentation is not very well and not very popular for the industry? (I donnu the game industries engine..btw) and just 1 or 2 active developer for that....I still haven't dig deeply into TROCS, but I hope it's better than Gamekit for this situation.
