Oxides catalyze some very interesting alkane reactions but tend to do it poorly. In this talk I describe computational work whose goal is to propose various chemical modifications of oxide surfaces aimed at making them better catalysts.  I will focus particularly on cation doping.  I will present calculations that show that lower-valence dopants improve the oxidizing power of a host oxide by activating surface oxygen atoms, while higher-valence dopants improve the oxidizing power by activating chemisorbed oxygen.  The calculations performed so far suggest a general rule: a Lewis acid on the surface facilitates strongly the binding of a Lewis base, and vice versa.  This rule explains a large number of computational results, has predictive power, and seems to be supported by experiments.  We explore the activation of alkanes by doped oxides by calculating the activation energy for breaking the C-H bond.  We find an interesting and general mechanism through which oxides perform this reaction.