??so what about "structural" / "axiomatic" (or something ...) description of cartanian geometry modeled on particular g/h ??? ...

??seems like g-torsoroid that's h-flat and k-flat _is_ [h sup k]-flat ... ???so does that screw up ideas about "natural geodesics" ??? ... ... and so forth ...

??what about ... ??getting graded nilpotent lie algebra by ... somehow ... ???starting with "bi-cartanian geometry" ... meaning reciprocal relationship between kleinian geometries sharing same stabilizer rep on tangent space ... ???or something .... ???relationship to "associated graded ..." ??? ??or something ??? ....

??what about "composition of rolling relationships" ??? ... or something ... ??relationship to some sort of span composition ??? ... ???? ... and so forth ... ??...

??so what _about_ "rolling one geometry on another" vs creating a new geometry this way ... ???or something ??? ....

??what about something about ... ??g-torsoroid that's h-flat, where downward normalization of f is non-trivial?? ... or something ... ??as peculiar because of ... ??usual bit about "making the g/h picture viable" as weird ... ??for example consider case g = h ... ?? ...???...

??so what about ... ??for example, non-geodesic light-like curves in ordinary "flat conformal space-time" (or something...) vs ... ??non-[conformally flat] spacetime ... ?? ... ??or something ... ???...

??so what about trying to use 1+1 penrose diagram to understand 2+1 flat conformal spacetime by "spacial radial symmetry" idea ?? ...