When dealing with large numbers of cells in biophysical simulations, it is important to properly manage the different substances that diffuse and react in and around cells. Although in an object-oriented programming environment it seems more natural to define cells as the basic objects, it turns out that individual substances are better suited to take this role. Here we describe the biophysical problem and our computational solution, and display the results obtained with a toy model. We find that the new implementation does not decrease performance and yet it leads to a much better structured and modular code. This will make more realistic programs with many molecular pathways much more modular and readily extendible.
