Abstract
Ontologies are being widely used in online science activities, or e-Science, most notably in roles related to managing and integrating data resources and workflows. We suggest this use has focused on enabling e-Science infrastructures to operate more efficiently, but has had less emphasis on scientific knowledge innovation. A greater focus on online innovation can be achieved through more explicit representation of scientific artifacts such as theories and models, and more online tools to enable scientists to directly generate and test such representations. This should lead to routine use of ontologies by scientists, and foster new and potentially different scientific results to help usher in next generation e-Science.