[ODE] Stability as a constraint?

[Gary R. Van Sickle]

Typing out loud here:

I'm probably either reinventing the wheel here or reinventing perpetual motion,
but this has been brewing in my mind for a while and I thought I'd vet it past
you knowledgeble folks.  Can stability be explicitly made a constraint, and
hence solve everybodys' problems once and for all time in the rigid body
simulation world?

By stability, I mean essentially "non-explosiveness", i.e. total system energy
is constant and non-infinite unless explicitly changed by the user (and hence
the energy change, hence new total energy, is still known and constant over the
integration interval).  The general scheme I'm envisioning is this:

1.  The World would have a global, scalar SystemEnergy variable.  This would be
the constant spoken of above.
2.  Each object would have a scalar Energy variable.
3.  Any time two objects interact, they'd exchange energy according to
conservation of energy laws.
4.  Each object's linear and angular velocities would be scaled or otherwise
constrained to the accumulated Energy.  This would for one thing eliminate the
"object spinning in space goes nutzoid" issue.
5.  All the Energy's would be constrained in some magical way to sum to the
SystemEnergy.  This would prevent any explosions.

Crazy?

--
Gary R. Van Sickle
Brewer.  Patriot.