Error in Book's Answer (5.5.9)

Michael W. Berry (berry@cs.utk.edu)
Thu, 29 Feb 1996 07:05:19 -0500


From: James Michael Hammond <hammond@cs.utk.edu>
Received: by adhara.cs.utk.edu (cf v2.11c-UTK)
id TAA16286; Wed, 28 Feb 1996 19:32:04 -0500
Date: Wed, 28 Feb 1996 19:32:04 -0500
Message-Id: <199602290032.TAA16286@adhara.cs.utk.edu>
To: berry@cs.utk.edu
Subject: back of the book might be wrong

Problem 5.5.9's answer in the back of the book sure looks like
it's wrong. When I crank it out by hand I get the answers

alpha = sqrt(5/7), a=c=7/25, b=8/75.

This is based on the assumption that the only place that the x^2 gets
into the act is when you are calculating the values for the solution
vector, in other words that you're solving the system

    /                \        /   \
    |	1    1    1  |   a    |2/3|
    |  -@    0    @  |   b =  | 0 |
    |  @^2   0   @^2 |   c    |2/5|    (where @ = alpha)
    | -@^3   0   @^3 |        | 0 |
    |  @^4   0   @^4 |        |2/7|
    \                /        \   /

if this is the case, then I'm pretty darn sure of my answer because it
verifies out for everything from f(x)=1 through x^5.

--JMike