[SciPy-user] Runge-Kutta ODE integrator in SciPy, odepack problems on SciPy Superpack for OSX?

[Anne Archibald]

On 17/03/2008, Nathan Bell <wnbell at gmail.com> wrote:
> On Mon, Mar 17, 2008 at 12:00 AM, Anne Archibald
>  <peridot.faceted at gmail.com> wrote:
>  >  I have to disagree with this. Getting numerical code right, and
>  >  robust, and bug-free, requires a lot of knowledge, skill and care.
>
> As does wrapping libraries.  Keep in mind that only a minority of
>  SciPy users/contributors will ever know SWIG or f2py, etc.  Keeping
>  things in Python greatly enhances accessibility of the implementation.

Yes, well, as the OP made it fairly clear that they just needed an ODE
solver, not specifically an RK solver, I had in mind using one of the
several good solvers built into scipy.

We could argue about the relative difficulties of wrapping libraries
versus implementing subtle numerical code, but I don't think that's
very productive. Both wrapping a library and implementing numerical
code are reinventing the wheel compared to using what's already
available (if it does what you want); whether you find it easier to
track down subtle numerical bugs or failures in a wrapper tool will
depend on your case.

Anne