Epilepsy is a major public health issue, not least because of the aging population in many developed nations and the known increase in the frequency of epilepsy and seizures in later life. Despite the massive scale of the problem and much research, epilepsy remains poorly understood. Despite more than 20 approved drugs in the developed nations and several non-pharmacological options, up to 30% of patients are still refractory to treatment. Despite over a century of pharmacotherapy and neuroscience research, rational design of anti-epileptic drugs (AEDs) is only now starting to yield results, because of the heterogeneity of the disease and our still limited understanding of it. Discovery and development of AEDs has been especially difficult, because of the regulatory issues of satisfactorily proving safety and efficacy, ethical constraints on placebo-controlled trial designs, the fact that seizures are typically widely spaced in time, and the fact that the person undergoing the seizure is typically in no state to remember, let alone assess, what happened. Several non-pharmacological therapies have been developed: brain surgery was first used more than a century ago; the ketogenic diet was first developed 80 years ago; and the vagus nerve stimulator was introduced recently. Pharmacotherapy remains the mainstay of treatment and is effective in most patients. AEDs can be roughly divided according to their time on the market. The first generation extends from the bromides and the barbiturates (the first of which was phenobarbital), to sodium valproate and carbamazepine. The second generation begins with felbamate and includes drugs approved from 1993 to 2000. "Next generation" drugs are still in clinical development and may reach the marketplace in the near future. Intensive research is being conducted both by pharmaceutical and biotech companies and by academic scientists and clinicians; our understanding of the condition is advancing rapidly but many challenges remain in discovering and developing better AEDs.