Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!sdd.hp.com!zaphod.mps.ohio-state.edu!rpi!dali.cs.montana.edu!milton!pezely@udel.edu From: pezely@udel.edu (Cowboy Dan) Newsgroups: sci.virtual-worlds Subject: VR system designs and use of Lisp Message-ID: <15641@milton.u.washington.edu> Date: 1 Feb 91 20:02:42 GMT Sender: hlab@milton.u.washington.edu Organization: HITL, R+D Cowboys (EntityOS Group at U of Delaware) Lines: 51 Approved: cyberoid@milton.u.washington.edu Below is a brief description of a system which uses, among other things, Lisp to make for a flexible design. Some people may object to using Common Lisp for VR systems at all. I am not saying that this is the only way to do it; just give it a chance and see if it works. Yes, there are more issues raised, such as run-time efficiency, but I think we should try to use existing languages before creating new ones. By creative use of data structures and object-oriented designs, you may find that all the elements you desire can be incorporated into an existing language. As I said before, I suggest using Lisp on top of another system platform. This top-level is an interface to the system, a system which would probably be written in some version of C. Then, you off-load all of the variable elements, such as i/o, to the top-level and make the system platform more like an operating system. And such designs look very much like the distributed OS designs from Tanenbaum's group, Amobea, and from AT&T- Bell Labs, Plan 9. As we (HITL and people at udel) develop more of our system and can release more information, the reasons for using common lisp will be clearer. This is, of course, not the only approach to VR, but it is a plausible one. At this stage of the game, I think VR systems will be mostly used for research and development of future systems. Of course, it would be very nice to have wide-spread VR and VR for the masses, but the realization of that might take some time. Some argue that we should be developing for today's low-end hardware; others argue that we should develop for high-end and future hardware, thus to alter one direction of hardware development. For now, I personally plan on developing for the high-end, and once such a system has been implemented, tested, basic applications developed, and benchmarks established, then at that point, I will focus on VR for the masses. I don't know if this is right or wrong, but I feel that more work can be accomplished with my talents and the resources made available to me by taking this approach. Comments? (send flames via email and save bandwidth) -Dan --