A short review of the hydrogen-oxygen reaction on Pt(111) is given in which the emphasis is put on the apparent contradictions between experiments reported in the literature. We consider these results in the light of our recent observation that certain defects appear to be very active sites for this reaction. Making the basic assumption that the reactive sites play the central role in the hydrogen-oxygen reaction, it is concluded that two fundamentally different mechanisms for the overall reaction are possible: (1) A process in which diffusion of oxygen to the reactive sites is essential for the reaction, and (2) an autocatalytic process in which hydrogen reacts at the reactive sites and is transferred to oxygen on the terrace via molecular reactions between adsorbed H2O, the reaction intermediates and unreacted oxygen. Based on these ideas, a model for the hydrogen-oxygen reaction on Pt(111) is proposed. With the model a consistent explanation of the reported measurements can be given, i.e., it can resolve the apparent contradictions between the different experiments. A transition between these mechanisms is expected in the temperature range where H2O starts to desorb and adsorbed oxygen atoms become mobile, i.e., around 250 K. We present experiments obtained in this temperature range using molecular beam titration and helium diffraction. The measurements clearly show the transition between the autocatalytic and the diffusion controlled mechanism and they are considered, therefore, as a strong support for the model.