Only about 1% of newly developed drugs are for tropical diseases, such as African sleeping sickness and dengue fever. While patent incentives have driven commercial pharmaceutical companies to make Western health care the envy of the world, the commercial model only works if companies can sell enough patented products to cover their R&D costs and produce profits for shareholders. The model thus fails in the developing world, where few patients can afford to pay patented prices for drugs. The solution to this devastating problem, say Stephen Maurer, Arti Rai, and Andrej Sali in the premier open-access medical journal PLoS Medicine, is to adopt an "open source" approach to discovering new drugs for neglected diseases. They call their approach the Tropical Diseases Initiative (www.tropicaldisease.org), or TDI. "We envisage TDI as a decentralized, Web-based, community-wide effort where scientists from laboratories, universities, institutes, and corporations can work together for a common cause." What would open-source drug discovery look like? "As with current software collaborations, we propose a website where volunteers could search and annotate shared databases. Individual pages would host tasks such as searching for new targets, finding chemicals to attack known targets, and posting data from related chemistry and biology experiments. Volunteers could use chat rooms and bulletin boards to announce discoveries and debate future research directions. Over time, the most dedicated and proficient volunteers would become leaders." The key to TDI's success, they argue, is that any discovery would be off patent. An open-source license would keep all discoveries freely available to researchers and--eventually--manufacturers. The absence of patents, and the use of volunteer staff, would contain the costs of drug development. Ten years ago, say the authors, TDI would not have been feasible. "The difference today is the vastly greater size and variety of chemical, biological, and medical databases; new software; and more powerful computers. Researchers can now identify promising protein targets and small sets of chemicals using computation alone."