The accurate identification of chemicals in text is important for many applications, including computer-assisted reconstruction of metabolic networks or retrieval of information about substances in drug development. But due to the diversity of naming conventions and traditions for such molecules, this task is highly complex and should be supported by computational tools.