numerical integration



James Amundson <amundson@users.sourceforge.net> writes:

> 2) There should probably be one higher-level interface to quadpack. I
> don't want to get rid of the current low-level interfaces, but I think a
> simpler starting point would be helpful.

There was some discussion about this in the `Quadpack checked in'
thread last year.  There was also a proposal for a common high-level
romberg or quanc8 like Maxima interface (which works nicely for
iterated integrals), with a prototype implementation for qags.

At that point, Raymond Toy wrote:

> Would this allow additional args to the integrator?  I'd really like
> an interface somewhat like plot2d so the user can specify additional
> parameters, like relative error, absolute error, array limits, etc.
> Using globals for the defaults is good, but allowing the user to
> override it in a call, is quite nice.

and I planned to reply (but didn't):

> There's no problem to pass options as long as we can distinguish the
> 3-argument call + n options from the 4-argument call + (n-1) options
> form.  So options in the style of the plotting functions are fine.
> 
> So I guess the idea is that the user may specify something like
> [absolute_error, 1d-8] without worrying about the actual name or value
> of the corresponding variable (or whatever incantation), which depends
> on the algorithm.

Then I thought "Hmm..." and waited for somebody else to do this.

> 1) The quadpack routines are entirely undocumented. One has to read
> numerical/slatec/quadpack.lisp to find them.

You don't say ;-)

Wolfgang