In light of anticipated changes in mathematics education, an alternative for the well- known "research-development-diffusion" model is presented. It is based on an integration of curriculum research and design embedded in "educational development." In this context curriculum development is described as purposeful and sensible tinker- ing. It is argued that the theory production implied in this process may be exploited in "developmental research." However, developmental research is not yet well established as a research discipline. The core of this article is an effort to explicate the characteris- tics of developmental research and to discuss its methodological aspects.