Michel Talon wrote:
> Raymond Toy wrote:
>
>
>> Ray, who thinks there's a reason why there are apparently no other
>> competing packages for LAPACK.
>>
>>
>
> Indeed numerical analysis is a hard subject! By the way, since the OP was
> speaking about numerical solutions of ODE, there is a wonderful program in
>
f2cl includes the ODEPACK from netlib, as a test case. Most of the
tests cases in ODEPACK work. Don't know if ODEPACK will solve the ODEs
needed.
> netlib, called colnew who is able to solve all sorts of boundary problems
> for ODEs. I have used it for my work and it really does wonders. But this is
> a very complicated program, with a complicated interface. I have tried once
> to run it through f2cl, but i think i remember it did not pass entirely
> through. Anyways it would be a great addition to maxima if someone more
> competent than me was able to make it work here, since the only other place
> it is ported to, as far a i know, is scilab, where the implementation is
> incomplete (there is no provision for continuations - a fundamental feature
> of this program). It can be found at:
> http://www.netlib.org/ode/colnew.f
>
I'll take a look at it. A quick 5 second glance at the code doesn't
indicate anything that f2cl should not be able to handle, except for the
Hollerith strings in FORMAT statements, which will have to be converted
by hand to strings, and for an ENTRY function that f2cl can sometimes
handle.
Ray