Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!cbosgd!ihnp4!qantel!dual!lll-crg!seismo!brl-tgr!gwyn
From: gwyn@brl-tgr.ARPA (Doug Gwyn )
Newsgroups: net.graphics
Subject: Re: ray casting
Message-ID: <1529@brl-tgr.ARPA>
Date: Mon, 16-Sep-85 16:31:41 EDT
Article-I.D.: brl-tgr.1529
Posted: Mon Sep 16 16:31:41 1985
Date-Received: Thu, 19-Sep-85 06:33:48 EDT
References: <1858@bmcg.UUCP>
Organization: Ballistic Research Lab
Lines: 21

>   The problem is I can't get a handle on how objects are described when they're
>   ray traced.

The description of 3-D objects is pretty much independent of ray tracing.
Two common techniques are "combinatorial geometry" (typified by PADL-2),
wherein solid primitives are combined with set operations (intersection,
union, difference) to make up a composite object, and "boundary
representation" (actually several different techniques), wherein the
surface of an object is represented as a collection of surface patches,
each of which is simply described (e.g. as a parametric bicubic).  There
are some more sophisticated approaches, too.

Just how you trace rays depends on how you represent the object.  In the
case of combinatorial geometry, the simplest approach is to intersect the
ray with every primitive solid (cylinder, cube, etc.) and perform Boolean
operations on the pieces of the ray resulting from each primitive.  In the
case of a surface description, one might be able to simply solve the
simultaneous equations for the ray and for the surface to determine the
point of intersection.  For more details, see Foley & Van Dam or some of
the more tutorial articles in past issues of IEEE Computer Graphics &
Applications.