[ODE] Reason for GeomTransformGroup not being core?

[Megan Fox]

The GeomTransformGroup contrib strikes me as a very useful thing to
have around, though I'm worried that it isn't core - does it have
certain know problems or glitches?


And this is an unrelated question, but - is there any reason to not
have a large number of spaces in a sim?  I mean, take my architecture
geometry... it's made up of wall chunks and floor chunks, each
separate.  Would it be any problem at all if I ended up having a space
for each and every chunk of geometry in the world (walls, NPCs, trees,
whatever), each containing typically one geom or possibly more in the
case of ragdolls/complex tables/etc?

As it stands, I need some way of clearly specifying what geometry
belongs to this chunk vs that entity (and which static geometry needs
to move when wall_chunk_01 gets a setposition command during level
building) - and spaces seem excellent for that, assuming there's no
huge perf cost associated with it?

(I would, of course, be using simple spaces for these sub-spaces,
given that they're tiny, containing only small numbers of
tightly-packed geometry)