Flow of observables

Suppose a classical Hamiltonian system (M,ω) with canonical coordinates. A function fC(M) has associated a Hamiltonian vector field Xf with a flow ϕs. Now, we can define a family of transformations in the algebra of smooth functions (i.e., observables)

ϕ~s:C(M)C(M)

by means of

gϕ~s(g)=gϕs

This is completely related to the Heisenberg vs Schrodinger picture discussion. The flow ϕs in M is some kind of evolution of states, like in the Schrodinger picture. And the flow ϕ~s corresponds to Heisenberg picture.

Observe that

ddsϕ~s(g)=dds(gϕs)=dgϕs()Xf=Xf(gϕs)={ϕ~s(g),f}

and

ϕ~0(g)=g.

And also, ϕ~s respect the Poisson algebra structure.

Reciprocally, if we start with a function f and we are able to associate a family of transformations {ϕ~sf} from C(M) to C(M) that preserve the structure of Poisson algebra and such that

ϕ~sf(g)=ϕ~s(g)

We proceed by taking the vector field {,f} and the associated flow ϕs, and using the two required properties, above. Since

ddsϕ~s(g)=ddsϕ~sf(g)

and

ϕ~0(g)=ϕ~0f(g)

we can show that ϕ~s(g)=ϕ~sf(g) by fixing a point PM.
We indeed have a duality, between transformations in M and transformation in C(M). Why do we pass from the flow in the phase space M to the "flow of observables"? Because this way, I guess, we can formulate everything in terms of algebras, and we get a deeper insight in the connection with Quantum Mechanics. Compare this with the discussion of derivations and 1-parameter group of automorphisms of the matrix algebra.