Re: A few (Octave) ideas

[Thomas Treichl]

Hi Jake,

I'll give my answers right after your questions:

> I'm aware of the OdePkg package on Octave-Forge, but this only
> implements three of the previously mentioned functions, ode23, ode45,
> and odeset.  I'm also aware of Octave's built-in function, lsode.
> These are all great functions, but I'm very interested in endorsing
> Octave over MATLAB in academia and as such, it's important that
> Octave has implementations of most of MATLAB's core functions.

For Odepkg, you also must say that we also have ode54 and my most 
loveliest solver ode78 and many many others that give very very good 
results... Isn't Matlab missing these solvers?

> Naturally, I've got a couple of questions:
> 1. Is there a reason why these functions haven't already been
> implemented? Lack of information on the algorithms, perhaps?

Implementing these functions would be a very good idea, I think Octave 
users would appreciate this work.

For OdePkg, I never had the goal to make a 1:1 implementation of all the 
solvers that are also available in Matlab. It was more important for me 
to solve all the different types of equations - this also resulted in 
different types of solvers (with a Matlab-like interface).

> 2. Is the maintainer of OdePkg (Thomas Treichl) already working on
> implementing these functions? I'd like to help, if possible.

No, I'm not. I must say that I've done what I can do but I think that 
big extensions from me won't follow anymore...

At the beginning of OdePkg my need was to have my solvers available in 
Matlab and Octave. This meanwhile has changed. If you would port the 
functions from Matlab-style to Octave-style coding (like John mentioned 
in his mail) then you could use all that work for the Octave core as a 
base and you don't need to start from the very beginning once again...

> 3. Are there any reasons why I shouldn't try to implement these functions?

Not from my limited point of view.

Best regards

  Thomas