From: broehl@sunee.waterloo.edu (Bernie Roehl) Subject: Re: TECH: My standard is better than your standard. Date: Thu, 16 Jul 1992 19:14:27 GMT Message-ID: Organization: University of Waterloo In article <1992Jul15.234125.7342@u.washington.edu> lance@unix386.Convergent.COM (Lance Norskog) writes: >Before this recitative, I should state that I'm positing a >[...] system in which decks are connected by an unreliable low-speed >network. Given that assumption, your conclusions make sense. However, my feeling is that network bandwidth will continue to increase, and that the problems of different decks getting out of synch are too serious to ignore. How does everyone else feel about this? It's interesting that we now have three divergent viewpoints: - highly centralized (mainframes or clusters of mainframes that maintain all the world state information - completely decentralized, where each 'deck' operates basically on its own with only minimal information exchange with others and all objects are implemented in the deck - a distributed model, where objects are implemented on a number of decks >Each deck will have to do a complete collision detection for the >current room. Reports from other decks will necessarily be out of >synch. That means that right from the word go each of us has a divergent reality; this (to me) is not a 'shared' virtual world as such. >Why? As long as their perceptions of the scene are approximately the >same, why do they all have to see exactly the same pixels? Pixels don't matter; however, if you see a door as open and I see the same door as closed, we aren't in the same world. >Why does it matter that you and I see the same dents? >We are merely going to notice "Ha! THX1138 is more dented than before". It comes down to what level of "sharing" we want in the reality. If you and I don't see the same things, it's hard for us to communicate in any meaningful way about the changing state of the world. "Boy, THX1138 sure got banged up -- look at that big dent in the door"; "There's no dent in the door"; "Yes there is"; "No there isn't" ad nauseum until we finally agree to disagree about what happened. >The master has to allow slaves some dynamic latitude in deciding what >to do until they get newer instructions. Controlling action at a >distance isn't easy. But it at least allows us to share a single, consistent world; if your copy of THX1138 and mine look different and bounce off in different directions, then there are two cars -- not one. More to the point, if my copy of myself behaves differently in your world and mine, then I'm not really in your world at all -- a clone of me is. -- Bernie Roehl, University of Waterloo Electrical Engineering Dept Mail: broehl@sunee.waterloo.edu OR broehl@sunee.UWaterloo.ca BangPath: uunet!watmath!sunee!broehl Voice: (519) 885-1211 x 2607 [work]