Well, I managed to make it work well. My doubt is about the answers I received from the method. I used the matrix found in wikipedia. Here it is:

S = | 4 -30 60 -35 |
| -30 300 -675 420 |
| 60 -675 1620 -1050 |
| -35 420 -1050 700 |

The results the wikipedia post as corrects can be found at the bottom of this page: http://en.wikipedia.org/wiki/Jacobi_eigenvalue_algorithm

My results are pretty the same. The only difference is in EigenVector 2 (E2) where my results has the same absolute value, but a different signal, i.e., where their result is negative mine is positive, and where their result is positive mine is negative.

Is it ok to happen?

Thanks for listening

Dear Petro:

Yes! It is ok. An eigenvector multiplied by any scalar, including -1, is still the same eigenvector. Convince yourself by plotting a three-dimensional eigenvector as a point in three-dimensional space. Draw a line through the origin. Now multiply the eigenvector by -1. Draw that point in the 3-space. It will reside on the line on the opposite side from the origin, symmetric to the first point. Thus, the same eigenvector. The same is true for any n-dimensional eigenvector plotted in n-space.

Ward Ciac