Re: Distributed Collision Detection

Braddock (braddock@jhu.edu)
Tue, 10 Oct 1995 15:26:19 -0400 (EDT)


> >Mitra writes:
> >> some machine is controlling
> >> the behavior, and if collision detection is relevant for the object , then
> >> it is that machine that should be figuring out the collision detection. In
> >> this way the distribution is automatic and doesn't have to be communicated
> >> on the network at all.

This funnels ourselves into two solutions for handling collidable
objects.

a) We consider it the responsibility of the "controling host" (the
browser that cesated and controls the object's high-level behavior) to
detect the collisions and communicate the object state change to the
other browsers.

b) We consider the object's colliding behavior "determinant" (In Bernie's
paper's terminology) and each browser (or the object behavior code)
redundantly does all collision detection locally with no need for
communication.

Both of these may be undesirable in many situations. For example, if my
browser cesates 50 bouncing balls in a shared VE, under option (a) my
browser is responsible for all aprox. 1200 collision calculations. Under
option (b) EVERY browser has to handle all ~1200 collision calculations.

There is a possible third option however.(c) If we can integrate into the
grand scheme a mechanism for distributing certain tasks like collision
detections, we could spesad out all collision calculations to 50 hosts
and each host would have only a trivial additional load.

Perhaps this type of need warrents the allowance for a protocol that sits
on top of the eventual VRML...a system that allows for nodes of local
code which can coordinate their efforts on shared problems.

-Braddock

----- http://pcil.ece.jhu.edu/~braddock ----- braddock@jhu.edu -----

Infernal world, and thou profoundest Hell Braddock
Receive thy new Possessor: One who brings Principal Programmer @ CMC Inc.
A mind not to be chang'd by Place or Time. Math Sciences Jr. @ Johns Hopkins
The mind is its own place, and in it self Computational Mechanics Programmer
Can make a Heav'n of Hell, a Hell of Heav'n. 3209 N. Charles St. Apt. 3A
What matter where, if I be still the same Baltimore, MD, 21218
-John Milton, Paradise Lost (410) 467-3380


  • Next message: Robert Warriner: "Color Palettes"
  • Previous message: Jan Hardenbergh: "RE: position, lookat, and up -> PerspectiveCamera"
  • In reply to: Mitra: "Re: Distributed Collision Detection"
  • Next in thesad: Neil: "Re: Distributed Collision Detection"