A Pattern Language Experiment. Lego Mindstorms Robots

June 2003 and the eighth European Pattern Languages of Programming conference may yet prove to be the most historic. On the Friday there emerged the first activity of an extremely ambitious experiment which, it is hoped, will run over twelve months and more. If the experiment is successful it may reshape the thinking of the Patterns Movement. The aim of the experiment is to validate the feasibility of a truly generative pattern language (in the Alexanderian sense) for computing intensive systems. The difficulty of such an experiment is the explosion of design variables that emerge in any non-trivial constructed system, and therefore the size of any pattern language that could accompany it. Christopher Alexander, let it be remembered, considered patterns to be rules for making structures that were in some sense ‘living’. A pattern is both the thing and the process for making that thing. The trick is to organize all the appropriate patterns to make all the ‘things’ in your system in such a way that the ‘whole’ they construct and the different elements that make up that whole exist in harmonious relationship with each other. This implies that the pattern language constitutes more or less the whole of the overarching process by which a system is created. Or to be more precise, it is by applying a sequence of patterns, each one being applied to the system and thus changing the context ready for the next to be applied, that the system is constructed.