Date: Jul 4, 2013 1:16 PM
Author: Nasser Abbasi
Subject: Re: An independent integration test suite
On 7/4/2013 11:49 AM, clicliclic@freenet.de wrote:

>However, it looks

> like example 6 from Chapter 7 is left unevaluated by Rubi. Am I right?

>

Good point. Yes, Rubi 4.1 did fully evaluate it, there was an "Int"

in the result, second term.

http://www.12000.org/my_notes/timofeev_integrals/rubi4/timofeev_rubi4.html

> I am not much surprised by Rubi's failure on numbers 10 and 11 from

> Chapter 7; Derive 6.10 cannot do them either. It may be acceptable to do

> nothing about this, I think.

>

> Martin.

>

> PS: Note that the 'model' results in my Derive files may not always be

> identical (or equal or equivalent) to the results given in Timofeev's

> book. In some cases, the integrands and/or results had to be corrected

> for misprints or mistakes. I may also have 'continuitized' some of the

> evaluations (effectively adding piecewise constants to make them

> continuous on the real axis).

>

Ok. What I did is just do

FullSimplify[mathematicaResult - bookResult ]

and hope to see either a zero or constant for the difference.

I assumed that a difference that still was a function of x would not

be accepted or that Mathematica could not simplify the difference.

In that case, I plot the difference to see if it is linear in x or not.

--Nasser