In this perspective the current state of photo-catalytic hydrogen production is considered, focusing especially on photo-reforming reactions. There are fundamental thermodynamic and kinetic limitations to direct water splitting and a number of strategies have been developed to overcome them. These include the development of catalysts which have little activity for the back reaction of hydrogen oxidation, systems with spatially separated H-2 and O-2 production and the interception of the oxygen species by sacrificial water reductants. Success has been achieved in these areas over the last 20 years or so, but more needs to be done in terms of the development of high quantum efficient photo-absorbers which operate over the useable light range, and of two-photon efficient processes which can extend the useable range further. There is also an urgent need for fundamental investigations of the nature of the range of surface processes which occur during photocatalysis, and which have to occur in synchrony. '