Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!wuarchive!texbell!texsun!csccat!larry From: larry@csccat.UUCP (Larry Spence) Newsgroups: comp.graphics Subject: Re: Solution to quartic eqn? Message-ID: <3375@csccat.UUCP> Date: 9 Nov 89 20:00:50 GMT References: Reply-To: larry@csccat.UUCP (Larry Spence) Organization: Computer Support Corporation. Dallas,Texas Lines: 17 In article CMH117%PSUVM.BITNET@VMA.CC.CMU.EDU (Charles Hannum) writes: >I didn't ask the question, but thanks for your input. However, Ferrari's >theorem yields a fast and very accurate answer. ^^^^ Are you sure about this? If it's the same closed-form solution that you find in the CRC books, etc., doesn't it use trig functions and cube roots? Seems to me there was a paper by Kajiya in the early '80s on numerical ray tracing, and there have been several in the last few years. My advice would be to go look at SIGGRAPH proceedings from 1981 on. Certainly, a closed form solution like the one suggested wouldn't take advantage of any coherence in the problem, unless you wrote all the trig stuff yourself. -- Larry Spence larry@csccat ...{texbell,texsun,attctc}!csccat!larry