Increasing user control over the environment has already started. New beta versions of the program allow the user to change the now static values of such application parameters as the spike mesh height field and the play speed. More menu items are being added that call special dialogs to let the user customize the application. As mentioned previously, the user should be able to set the bounds for the recording interactively at the during the file selection. The user will have interactive control over the material attributes of the meshes in the simulation. It has been suggested that further control be given over the relative positions of the visible meshes in the viewing area. An algorithm may be developed to arrange the visible populations according to the presence of other populations in the scene graph. This would condense the viewing area allowing for closer views of all the interesting information. The ability to save certain information in data files would be helpful. Even saving the contents of the entire recording data structure to a file might be helpful for the fast recall of a previously calculated simulation. Saving user preferences such as material attributes would make customization easier. The ability to capture the viewing area to an image file , or repeatedly capture subsequent frames to a video file would make demonstration of the simulation portable. As the application is tested in the Department of Biology at the University of North Carolina at Charlotte, we should be able to get a better idea of the priority of the changes that need to be made.

Another focus of future work will be to experiment with ways to increase the information represented at each time step. The most straight forward way is to increase the number of cell populations in the simulation. Preliminary experimentation suggests that an increase in the number of cell populations does not bind the graphics subsystems of the Indigo2 workstations that the simulation is being run upon.

Most of the Inventor scene graph has been optimized, but there remain areas for optimization in the implementation. The Inventor mesh object could possibly be rewritten so it does not make a copy of the values it uses for the mesh coordinates. The inventor mesh could also be replaced by triangles. A triangle representation would allow only the points that change to be updated rather than the present system of updating all of the z values in the meshes. If this were the case, the advantage of storing only the changing spike values for each population would be realized in speed increase as well as space saving.