Re: gauss chebychew quadrature in integration package

[Ben Abbott]

On Jun 22, 2008, at 3:59 PM, Thomas Plehn wrote:

> my whole octave session with all user input is shown on the picture  
> accessible under the posted link
>
> ----- Original Message ----- From: "Ben Abbott" <address@hidden>
> To: "Thomas Plehn" <address@hidden>
> Cc: <address@hidden>
> Sent: Sunday, June 22, 2008 9:14 PM
> Subject: Re: gauss chebychew quadrature in integration package
>
>
> > On Jun 22, 2008, at 2:54 PM, Thomas Plehn wrote:
> >
> > > http://img28.picoodle.com/img/img28/4/6/22/f_quadraturem_a5f90c3.png
> > >
> > > hello,
> > > I have tried to integrate a simple function in octave using the  
> > > integration
> > > package.
> > > the function is f(x,y)=x^2*sin(y)
> > > the integration limits for both x and y are from 0 to 1.
> > > the result should be (1-cos(1))/3 = 0,15323
> > > this result is yielded by quadndg, gquadnd
> > > but cquadnd yield a different result
> > >
> > > should this be considered as a bug?
> >
> > Can you post a bit more information. Perhaps a short script   
> > illustrating specifically what you did?
> >
> > Ben

ok,

I've not used the integration package before, but modified you example  
to create a simpler version.

> pkg load integration
> function y = f (x)
> y = 1;
> endfunction
> cquadnd ("f", 0, 1, 1)
ans =  1.5708

The answer is obviously wrong (should be 1), and as far as I can tell  
you are using cquadnd in manner consistent with the documentation.  
Thus, it does look like a bug to me.

To report the bug, please go to the link below

        https://lists.sourceforge.net/lists/listinfo/octave-dev

Ben