Reversible computation is a paradigm allowing computation to proceed not only in the usual, forward direction, but also backwards. Reversible computation has been studied in a variety of models, including sequential and concurrent programming languages, automata, process calculi, Turing machines, circuits, Petri nets, event structures, term rewriting, quantum computing, and others. Also, it has found applications in areas as different as low-power computing, debugging, simulation, robotics, database design, and biochemical modeling. Thus, while the broad idea of reversible computation is the same in all the areas, it has been interpreted and adapted to fit the various settings. The existing notions of reversible computation however have never been compared and categorized in detail. This work aims at being a first stepping stone towards a taxonomy of the approaches that co-exist under the term reversible computation. We hope that such a work will shed light on the relation among the various approaches.