Almost all viruses have polymerases which are suitable targets for antiviral drugs. The development of selective polymerase inhibitors started with screening of compounds in virus-infected cell cultures and the mechanism of action was investigated once an inhibitor had been found. Especially nucleoside analogs were screened as their triphosphates were potential substrates for polymerases. However, the stepwise phosphorylation by cellular, and sometimes viral, kinases to the active triphosphate prevented a truly rational design of polymerase inhibitors. Nucleotide analogs offers a type of compounds which could be designed in a more rational way than nucleoside analogs since the first, most selective, phosphorylation step is eliminated in the path to the active inhibitor. The development of pyrophosphate analogs made rational design possible since these compounds act directly on the viral enzyme, but the room for structural variation was limited. The non-nucleoside HIV reverse transcriptase inhibitors are direct inhibitors and can thus be designed in a truly rational way by use of structure information on the enzyme-inhibitor complex by use of X-ray and NMR. This rational design of allosteric inhibitors is also being used in the development of inhibitors to other viral polymerases.