We completely classify one-sided Markov chains up to measure-theoretic isomorphism. The classification is effective and computable.
