Floating-point calculation is inherently approximate. You might want to
read David Goldberg's "What Every Computer Scientist Should Know about
Floating-point Arithmetic"<http://www-users.math.umd.edu/~jkolesar/mait613/floating_point_math.pdf>;
for
discussion.
Maxima provides *arbitrary-precision* floating-point, which is also
approximate, but can be calculated to a large number of digits, for example:
z: (10+6*sqrt(3))^(1/3) + (10-6*sqrt(3))^(1/3);
bfloat(z),fpprec:1000 => 2.0b0 (accurate to 1000 decimal digits)
It would be nice if Maxima could determine that z is *precisely* 2 by
symbolic methods; I don't know how to do it or whether Maxima can do it --
perhaps others can help.
By the way, when sending questions and comments to the Maxima mailing list,
it is best to send a textual form (e.g. from string(z) ) rather than an
image, which requires retyping (which is annoying and error-prone). I'm
afraid I don't know how easy that is to do from wxMaxima.
-s
On Mon, Jul 25, 2011 at 10:15, Jan Hendrik M?ller <
jan.mueller at math.uni-dortmund.de> wrote:
> Dear all,
> I recently looked for integer values of cubic roots expressions. I was
> surprised that wxMaxima computed the second expression (of the attached
> screenshot) wrong, because both are integer. Can someone help me/explain the
> reason for the wrong result?
> Thanks a lot
> Jan M?ller
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>
>