Path: utzoo!attcan!telly!lethe!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!samsung!crackers!transfer!lectroid!jjmhome!smds!rh
From: rh@smds.UUCP (Richard Harter)
Newsgroups: comp.lang.pascal
Subject: Re: Roots of polynomials
Summary: Unwarranted and Unwelcome
Message-ID: <268@smds.UUCP>
Date: 7 Dec 90 10:06:18 GMT
References: <2173@mentor.cc.purdue.edu> <1990Dec5.175558.29859@dg-rtp.dg.com> <4529@idunno.Princeton.EDU>
Organization: SMDS Inc., Concord, MA
Lines: 46

In article <4529@idunno.Princeton.EDU>, subbarao@phoenix.Princeton.EDU (Kartik Subbarao) writes:
> In article <1990Dec5.175558.29859@dg-rtp.dg.com> hagins@gamecock.rtp.dg.com (Jody Hagins) writes:
> >In article <2173@mentor.cc.purdue.edu>, k3u@mentor.cc.purdue.edu (Barry Smith) writes:
> >|> Does anyone have the source code to routines that can extract
> >|> the complex roots of polynomials?  If there is a book that has
> >|> the source code that would be fine too.

	[Snide comments which make unwarranted assumptions about the
	request being one for meeting homework deleted.]

> Yes this is true. But why are you assuming that he wants to do his homework
> on USENET? Don't you think its a bit prejudging to assume right off the
> bat about a person's intentions? Especially the part about "slower"
> students and the like. That's getting downright condescending. I really 
> don't think that kind of tone belongs in an educational newsgroup (your own 
> words). What if he is doing some independent work on the stuff and wants to
> find out whether there already exists code to do what he wants? There
> certainly is no point to re-inventing the wheel. I certainly am not going
> to take too much time writing someting that I need done quickly and is
> already around.

Perhaps it is childish of me, but I was highly offended by Jody's comments,
which I felt were snide, unwarranted, and, considering the specific request,
silly.  First of all many people who are not computer science students
taking a numerical analysis course have reason to need a routine to find
the complex roots of a polynomial -- for example, physicists, electrical
engineers, chemists, mechanical engineers, etc.  It's a very common problem
in applied math in all fields.  Secondly it's a hard problem.  Quite frankly
I doubt that Jody has ever written a good general purpose routine for
finding complex roots of polynomials.  I would be surprised that anyone
who has had to deal seriously with the problem would dismiss it so casually
and comptemptuously.  Most of the methods given in the numerical analysis
texts fail badly on ill-conditioned polynomials.  

In response to the original query -- your best option, if it is available,
is to cast the problem as an eigenvalue problem and use a canned eigenvalue
package to get the roots.  Your system should have one available.  If not
I believe you can get the EISPACK package via anonymous ftp.  In any case
the best groups for this request are comp.lang.fortran, comp.math, and
comp.math.num-analysis (sp?).  Also check to see whether the IMSL routines
are available on your system.
-- 
Richard Harter, Software Maintenance and Development Systems, Inc.
Net address: jjmhome!smds!rh Phone: 508-369-7398 
US Mail: SMDS Inc., PO Box 555, Concord MA 01742
This sentence no verb.  This sentence short.  This signature done.