canonical quantization
Canonical quantization is the recipe that turns a classical Hamiltonian theory into a quantum one. It is the bridge that Dirac built when he noticed that the classical Poisson bracket and the quantum commutator obey the same algebra. The prescription is almost mechanical: take your classical system in canonical coordinates, promote the coordinates and momenta to operators, and replace Poisson brackets by commutators. Out comes quantum mechanics.
The core rule is the correspondence {f, g} -> (1 / i hbar)[F, G], where [F, G] = FG - GF is the commutator of the corresponding operators. Applied to the fundamental brackets {q_i, p_j} = delta_ij, it produces the canonical commutation relations [x_i, p_j] = i hbar delta_ij (with [x_i, x_j] = [p_i, p_j] = 0) -- the seed of the entire theory, and the source of the Heisenberg uncertainty principle. Classical observables become Hermitian operators on a Hilbert space; the classical Hamiltonian H(q, p) becomes the Hamiltonian operator; and the classical time-evolution law df/dt = {f, H} becomes the Heisenberg equation of motion dF/dt = (1 / i hbar)[F, H]. The formal parallel is exact and is precisely why the Hamiltonian formulation, not the Lagrangian, is the natural launch pad for quantization.
Canonical quantization is the standard route into quantum mechanics and quantum field theory, but it is a heuristic, not a theorem, and it has real ambiguities and limits. Because operators do not commute, a classical product like q p has no unique operator counterpart (should it be x_hat p_hat, p_hat x_hat, or the symmetric average?) -- this is the ordering ambiguity, resolved only by additional physical input. The Groenewold-van Hove theorem proves that no quantization can consistently map all classical observables to operators while preserving all brackets, so the correspondence is guaranteed to break for higher-order quantities. It also depends on a choice of canonical coordinates and does not straightforwardly respect general canonical transformations. It is a powerful, imperfect ladder from the classical to the quantum world.
Starting from the classical oscillator H = p^2/(2m) + (1/2) m omega^2 x^2, canonical quantization imposes [x, p] = i hbar and reinterprets H as an operator. Defining ladder operators from x and p then gives the quantized energies E_n = (n + 1/2) hbar omega -- the zero-point energy and evenly spaced levels emerging directly from the single commutator [x, p] = i hbar.
Canonical quantization of the classical oscillator, via [x, p] = i hbar, yields the quantized levels E_n = (n + 1/2) hbar omega.
Canonical quantization is a well-motivated heuristic, not a rigorous derivation. The ordering of non-commuting factors is ambiguous, the Groenewold-van Hove theorem forbids a perfect bracket-to-commutator map for all observables, and the procedure depends on the chosen canonical coordinates -- so it must be guided by experiment and consistency, not applied blindly.