I think you have all the spec mechanisms you need in X3D to make the subset your talking about.
Why? Because I don't think the language I am proposing will be "X3D". Since X3D is designed for web delivery of interactive content it seems appropriate to derive the language from X3D. However, it is likely to be sufficiently different that I doubt there would be any point in labelling it X3D.
Sounds like a big amount of wasted effort to start afresh and have to redefine everything. If your interested in working within the X3D spec then I'm game for helping out. If its creating yet another way to describe the same 3D graphics concepts, then count me out.Furthermore, since the aim is to create a very lean language for mobile simulation content delivery, it is best not to burden the language with any perception of an overweight X3D.
I'd like to know other peoples thoughts on the above,