Re: "Object oriented" description of functions

[Brian Gough <bjg at network-theory dot co dot uk>]

For the moment I would prefer to keep things as they are, just to
avoid extra work. To access a function make a new gsl_function struct
directly,

  gsl_function f;
  f.function = fdf.f;
  f.params = fdf.params;

Maybe we'll change this in the future.  I agree that current system could
be improved.