We continue to develop the code which will allow a machine to determine the optimum layout for a desired space. At this stage, our program chooses coordinates and rotations that are completely inside the finished room and away from the middle. It also iterates through hundreds of combinations to make sure none of the components intersect (overlap)--something generally impossible in physical space! Most of the work up till now is learning the SketchUp API so as to manipulate the model via Ruby code. Next step for us is to develop neural networks by which the computer will 'learn' to arrange components with no instructions other than available methods and a variety of success functions.