Re: [Help-gsl] Interpolation - moving from v 1.9 to 1.11

[Brian Gough]

At Sat, 12 Apr 2008 11:44:34 -0400,
Cathy Laporte wrote:
> I am currently in the process of migrating from gsl-1.9 to gsl-1.11.  I
> noticed that there was a change in the interpolation routines of the
> library.  Namely, the interp_init routine now checks that the x values
> of the data are sorted in increasing order.  This change has brought a
> bug in my own code to my attention: occasionally, it tries to create an
> interpolant with a small number of incorrectly ordered (but quite
> similar) x values at one extreme of the range of my data.  Using
> gsl-1.9, this bug was rather asymptomatic as the library did not issue
> any error messages and most of my interpolation queries were probably
> not in the dubious part of my x axis anyway.  Nonetheless, for the
> purpose of understanding what sort of impact this bug may have had on
> the output I previously obtained from my code, I would like to
> understand how interpolant queries behaved with incorrectly ordered x
> values in gsl-1.9.  Does anyone know?

I don't know exactly but there are two factors - (i) the region is
located with a binary search so it can change the results of that and
(ii) the behavior inside the region depends on the interpolation
method used.

-- 
Brian Gough