Vo, All these things are just the energy vector.
At the moment we are constrained to the wheel running on road. (It's been like for millennia until someone comes up with non-aerofoil levitation). The most graceful solution to mobility of apelike-bipeds is the electric motor >80% efficient. The whole business of heat and reciprocating engines will seems quaint in ten years. I am of the opinion that I can make a heat battery say CaO + water (or how about CaC + H2O -> ethylene for you ICE fans). Useless low grade heat at the moment but with the ferrofluid temporary remenance cycle I do at uni. heat will just routinely become another form of energy you can transform too at 100% efficiency. I mean a pendulum (in vacuo) cycles KE to PE and we see no mystery. Take a coupled ensemble of those pendulums and the nice motion you've set up at the start becomes randomised - you have trouble getting it back and we call that entropy increase. However the individual pendulums are still, perfectly inter-converting KE and PE. Put something on the scale of these little atomic pendulums (moving on in the argument) and you can tap and cohere this motion. Yes it a Maxwell Demon but it can be done if you know your way around the Maxwell Demon arguments. Basically you must have a discriminant function to decide what's fast and slow (a gate etc.). The arguments come down to that your function gets randomised because it is getting knocked about too (its small, on the scale, right?) or that it is blinded by the isotropy of the radiation, or that it must have a little computer running it and this inevitably rejects more heat into the environment than the d.f. is trying to reduce. All the old arguments are bogus if you use phase changes and think out of the square. This can be argued macroscopically (thermodynamics T-S diagrams, meso-scopically considering the chemical potential and nano-scopically by molecular dynamics simulations. That's what I'm up to and I feel a renewed vigour after all the years of boredom. It is an interesting subject.
