Re: [pdf-devel] PDF sampled functions and interpolation

[Hardy Falk]

Am Donnerstag, 13. September 2007 23:47 schrieb Karl Berry:
>       Brian Cough (maintainer of gsl) is worried about the
> efficiency of the gsl functions for a software such as the
> pdf library. I disagree with him. I think the gsl would fit
> our needs quite well.
>
> Surely it's far too early to worry about efficiency.  If gsl
> will move development forward by a leap and a bound, that
> seems all to the good. Down the road, if it turns out some
> function or another is bottleneck, it can probably be
> optimized on its own.
>
> Computers have lots of cycles available.  Human programmers
> don't :).
On the other hand, gsl would require a lot of glue.
I would not recommend gsl if the *only* reason is cubic spline 
interpolation. It is required for completeness, but xpdf seems 
to fall back to the default \Order 1.