The emergence of open-source libraries and development tools in the last decade has changed the process of academic software development in many ways. In medical image processing and visualization this change is especially evident, also because open source projects are actively furthered by grant funding institutions. This manuscript presents the use of such development tools and libraries at the UNC Neuro-Image Analysis Laboratory for open source applications and tools. We have also experienced in our research that the development of open source in academics raises the issue of access to unpublished methodology. The strategy at our laboratory is to combine all in-house libraries and applications into a single repository that consists of two parts: a fully open source part that is distributed under a Berkley-style license and a private, closed source part with unpublished tools and methods. Access to the open source part is unrestricted, whereas the private parts can only be downloaded via cvs user login. This setup solved our issues regarding unpublished methodology, as migration from the private to the open source part is very simple. Overall our experience with this development environment within the academic setting is very positive.