Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!sdd.hp.com!decwrl!elroy.jpl.nasa.gov!jarthur!mti!adrian
From: adrian@mti.mti.com (Adrian McCarthy)
Newsgroups: comp.graphics
Subject: Re: Where to publish ray tracing algorithm?
Message-ID: <1150@mti.mti.com>
Date: 6 Sep 90 16:07:01 GMT
References: <1145@mti.mti.com>
Reply-To: adrian@mti.UUCP (Adrian McCarthy)
Organization: Micro Technology, Anaheim, CA
Lines: 23

Thanks for the overwhelming support.  I've received more mail than I can
keep up with.  Several people are now looking at my algorithm, and I'm
actively researching all those references I've received.  I'm going to
wait until I get a little more feedback and can come up with a second
draft before I send my algorithm out to other folks.  The real early
responses seem to suggest that my idea holds water and is original.

Since so many people showed interest, and I can't keep up with all the mail,
here's a taste of the idea:  Rather than using recursive or hierarchical
spatial subdivision techniques to reduce ray-object intersection tests
(which are of O(log n) algorithms) many instances can use a surface map for
a single bouding volume as a lookup table to immediately determine a small
subset of objects to test (which is truly of O(1)). (Small subset here is
roughly equivalent to the set of objects in the smallest volume in a
comparable hierarchical scheme.)  It's *not* general, but the cases where it
is useful are many, especially if you create CSG objects of many primitives.
It can be combined with traditional methods.

Please don't be offended if I didn't respond to you directly.  I really was
overwhelmed by the response.

Thanks all,
Aid.