[boost] Re: Re: Re: Numerical Integration - interface

From: Dan Mcleran (dan.mcleran_at_[hidden])
Date: 2002-10-30 16:13:55

"Martin Weiser" <weiser_at_[hidden]> wrote in message
news:200210301421.46600.weiser_at_zib.de...
> On Montag, 28. Oktober 2002 22:24, Dan Mcleran wrote:
>
> > Martin,
> >
> > I've expanded on your initial thoughts and written some code that
> > performs a trapezoidal-rule integration on a couple of functors (1/x
> > and x^2).
>
> Dan,
>
> I've written an interface framework which I think is more flexible, and
> nevertheless easy to use. It's far from complete, of course, and the
> implementation is clearly suboptimal, but you get the idea.
>
> Let me know what you think.
>
> The code below compiles and runs with GCC 3.2. No other compiler was
> tested by now.

Martin,

This looks much better than what I had. The only thing I didn't like was
that the library always constructs an instace of TrapezoidalRule every time
integrate is called (or am I mistaken). The other down-side is that neither
Borland nor GCC 2.95.2 will compile your code.

I'm using your code as a starting point and making some mods. I would like
to show it to you when I get it working. I have to spend some time on my
real job today but hope to look at this some more tonight.

Thanks,

Dan McLeran