Abstract
We compare several methods for approximate implicitization by piecewise polynomials which have been developed by the authors, and a linear-algebra based numerical method for implicization which is provided as a part of MAPLE. We investigate both quantitative criteria (such as computing time, memory use, and the error of the approximation) and qualitative criteria. As demonstrated by the results, piecewise approx. implicitization is able to handle surfaces arising in industrial applications. However, special care has to be taken to avoid additional branches and unwanted singularities.