Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site uiucdcsp.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!ihnp4!inuxc!pur-ee!uiucdcsp!ashby From: ashby@uiucdcsp.UUCP Newsgroups: net.math Subject: Re: real roots of polynomial Message-ID: <9600003@uiucdcsp.UUCP> Date: Mon, 11-Feb-85 22:01:00 EST Article-I.D.: uiucdcsp.9600003 Posted: Mon Feb 11 22:01:00 1985 Date-Received: Wed, 13-Feb-85 04:12:11 EST References: <7942@brl-tgr.UUCP> Lines: 6 Nf-ID: #R:brl-tgr:-794200:uiucdcsp:9600003:000:263 Nf-From: uiucdcsp!ashby Feb 11 21:01:00 1985 A problem with simply coding a CRC formula is that it may be numerically unstable. I believe finding the roots of the polynomial as the eigenvalues of the companion matrix is numerically reliable as well as easy to code -- if you happen to have EISPACK around.