Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!tut.cis.ohio-state.edu!pt.cs.cmu.edu!andrew.cmu.edu!nl0s+
From: nl0s+@andrew.cmu.edu (Nathan James Loofbourrow)
Newsgroups: comp.graphics
Subject: Re: Solution to quartic eqn?
Message-ID: <UZKFA9W00WB8I9EjpR@andrew.cmu.edu>
Date: 9 Nov 89 06:02:49 GMT
Organization: Class of '92, Carnegie Mellon, Pittsburgh, PA
Lines: 13

If you are willing to settle for approximate answers (and in most graphic
applications, I would hope that you would :-) ), I would recommend using
some variation on Newton's Method for determining roots of functions.
If you separate your at-most-four roots by subdivision (of the interval in
which the roots can occur; you should be able to produce a maximum interval
without too much trouble, much as you might produce a bounding volume), you
should be able to approximate the roots to any desired precision.

Now no one is saying this is fast.

Nathan Loofbourrow
nl0s+@andrew.cmu.edu
nl0s@andrew.bitnet