On Fri, May 25, 2007 at 08:52:16PM +0200, Andrej Vodopivec wrote:
> On 5/25/07, Robert Dodier <robert.dodier at gmail.com> wrote:
> > On 5/24/07, Andrej Vodopivec <andrej.vodopivec at gmail.com> wrote:
> >
> > > Function names should be suggestive.
> >
> > Yes, nintegrate certainly is suggestive, isn't it?
...
> The name stands for numeric integratation. It is taken from
> Mathematica. When I said that I didn't want it to do whatever it can
> to get the best result I meant that it shouldn't use symbolic
> integration because it is not a numeric method.
basically, nintegrate implies an expert system to identify a method
which should be applicable to a particular problem, and carry that
method out.
This is convenient for a large number of problems which are generally
easy to numerically integrate. It is of course disastrous if you give
it a problem where the method chosen gives a wrong answer.
A routine called "nintegrate" should be rather conservative and
extremely robust against a wide variety of inputs. A routine called
qp_integrate or something similar can safely assume that the user is
aware that the method will be quadpack, and that the types of problems
and errors that will be encountered are those related to quadpack.
I think the concern among the group is just that the standards that
are implied by "nintegrate" are higher than the standards of quadpack.
--
Daniel Lakeland
dlakelan at street-artists.org
http://www.street-artists.org/~dlakelan