A 15:29 +0100 31/08/99, Higgo James a écrit: >It really is not difficult to replicat just those parts of the universe >necessary to fool the inhabitants into thinking it's real.
I think it is not difficult, it is impossible! The program just >has to 'kick back' in the right way. It does not need to model every >particle of every tree in every quad. > >But here we get back to our old stamping ground... exactly. First your remark makes sense only if the inhabitants are real, i.e. they are just interfaced with a fake reality (what is the case in Matrix I believe), e.g. by cutting all nerves and calculate the I/Os. It is different from the case where the inhabitants themselves are simulated. The problem is to know if it is possible to cut the reality at some level of description and stay completely coherent. I think not, because there are many phenomena that arise in fact (at least) from molecular structure and cannot be calculated without this information. This is particularly true within our present knowledge of the world, which is perfectly consistent with all measurements that we can make down to the particle level. I don't see how a computer could simulate that without taking explicitely into account this description. For example I know that I could take any small dust grain out of my environment and analyze it by a mass spectrometer or through X-ray diffraction, or Mossbauer spectroscopy, and that the results I would get should be perfectly consistent with each others. Gilles