Re: [Texmacs-dev] Problem of division in scheme.

[Joris van der Hoeven]

On Tue, Sep 29, 2009 at 10:16:36AM +0200, Corcelle wrote:
> 
> > There are differences in the arithmetic of guile between 1.8*
> > and before, due, among others, to the use of gmp.
> > It is better not to depend to much on the arithmetic of guile...
> > 
> > Best wishes, --Joris
> 
> Thank you for the precision.
> 
> I use a lot this arithmetic of scheme-guile in the math extension of the 
> graphics mode. In fact, all the constructions are based on complex operations.
> What would be the best solution to go around this problem ? Use another 
> "engine" for the calculus ? 

If you want to stick to scheme, which is probably simplest for what you want,
then you can write your own scheme extension for complex arithmetic.

> In that case, which one should I choose and would it be as easy as scheme to 
> use ?
> It's true that, in that case, I could have more powerful tools for my 
> calculations.

You may also use any of the TeXmacs plug-ins for doing the computations,
e.g. Maxima (or Mathemagix, in a relatively near future).
However, the communication is currently text-based, so not very suitable
for intensive computations (except if you write all your routines in the 
plug-in).

For the moment, Scheme is probably the best solution.
You may also experiment a bit with the goops,
an object oriented extension of Guile which will allow you
to overload mathematical operations, such as +, -, *, etc.
For the moment, I do not use it, but it might be worth to
investigate how robust it is.

Best wishes, --Joris