As far as making simulations. You can provide test operations I think, but freenet presents difficulties that can't be addressed in any simulation. I think it best to assume that the current use is the ultimate simulation and that everyone push the system as far is it will go. The larger the node set, the larger the complexities and the more likely the bugs.
I guess a question could be: At what rate of growth does probability of complexity errors grow as NUM_NODES goes to Infinity. Has this type of problem been addressed yet? Anyway, my solution is to keep tring to grow the node population as fast as possible to determine if there might be chaos error which never goes away anyway, but you must be aware of. You need to get feed back from the other nodes for this. I propose a tempory debug envoronment that will report errors to a central site or some freenet key. If complexity errors grow at log(n) where n is the number of nodes, then problems will be managable. However if error grows exponentially, the system is doomed in complexity theory. IMHO Jay Ferguson _______________________________________________ Devl mailing list Devl at freenetproject.org http://lists.freenetproject.org/mailman/listinfo/devl
