A method, which we suggest to call the Empirical Maximum Entropy method, is implicitly present at Maximum Entropy Empirical Likelihood method, as its special, non-parametric case. From this vantage point the entropy-based empirical approach to estimation is surveyed.