We discuss several methods to produce superpositions of optical coherent states (also known as cat states). Cat states have remarkable properties that could allow them to be powerful tools for quantum information processing and metrology. A number of
proposals for how one can produce cat states have appeared in the literature in recent years. We describe these proposals and present new simulation and analysis of them incorporating practical issues such as photon loss, detector inefficiency, and limited strength of nonlinear interactions. We also examine how each would perform in a realistic experiment.
