From: demodb@manchester.dab.ge.com
Subject: TECH: Simulators in VR
Date: 	Fri, 10 Jul 1992 08:25:12 -0400




     <  Going to an "intermediate" graphics box connected by a thin
     <  (low-bandwidth) wire would be a step backwards in many ways.
     <  Traditional visual simulation platforms have been of the
     <  big-black-box variety.  You load up a database and then over a
     <  relatively low-bandwidth channel, the host computer talks to the
     <  IG (image generator) sending down updates for the state of
     <  moving coordinate systems, animation sequences, etc.  Spatial
     <  inquiries (collision, line-of-sight) are also communicated over
     <  this link.

     <  This sounds like a nice, clean partitioning, but it is also
     <  tremendously restrictive.  What if you want to code your own
     <  dynamic geometry for surf crashing on a beach?  What if you want
     <  to interactively compute and visualize particle paths in a fluid
     <  flow?  What if you want to do interactive deformations of
     <  objects?  The separate host/IG model breaks down because the
     <  communication channel is too slow to transfer the geometry each
     <  frame and the black box doing graphics doesn't know about waves,
     <  particle advection or editing elastic surfaces.

 In the past, the simulator itself being tightly coupled to commercial/
 military aircraft development and DoD training programs has reflected
 those applications. But, in general, the CIG is a stand alone programmable
 piece of hardware (ie. doesn't need a "host" to work), that is capable
 of a much wider variety of functions then traditional roles dictate.
 One GE Compuscene has been doing double duty as both a ship bridge control
 trainer and as a visualization tool for performing realtime analysis
 of ship hull designs using a programmable dynamic sea state (waves)
 capability. Just as the workstation market has benefited from in-house
 and third party software development on their platforms to meet the need of
 customer requirements, so is the evolution of new applications occurring for
 CIG.
 It appears, from what I've observed on this news group and conference
 proceedings, is that the acceptance and proliferation of VR to the masses
 may be first seen in some form of interactive entertainment and/or DoD
 network simulation projects. This is not due to any perception of these
 applications being of greater importance than flow analysis or scientific
 visualization, but more of a funding issue. 
 The simulator, growing from it's history of proven applications involving
 database(worlds) development, complex collision detection, vehicle dynamics
 computer guide interaction, display tech., networked training, etc..., will
 prove to have a logical extension into presently attainable VR applications.

     Bob Hichborn
     Advanced Product Applications
 
     Phone 904-239-2704
     Fax   904-239-2924
     E-mail  demodb@sunny.dab.ge.com