To analyze the error obtained in the evaluation of the shape of the surface with the method that uses a linear array of sources, we performed a numerical simulation introducing Gaussian random errors in the coordinates of the centroids of the bright spots on the image plane, and in the coordinates of the sources (apertures of the mask). Here we found that to measure the quality of the surface with accuracy lower than 8 µm we have to place the sources with accuracy better than 0.5 mm, and measure the coordinates of the centroids with accuracy better than 0.5 pixels. Additionally, we performed an analysis of the aberrations effects on the evaluation due to a steep surface and the CCD camera lens. Here, we found that in the best image plane for all the image points the main aberration is astigmatism, which is symmetric and do not change the position of the centroid. Also, the analysis of the CCD camera lens was performed experimentally by the star test for different positions of the object in the visual field and for different apertures of the CCD diaphragm; here, in the worst case we observe changes in the positions of the centroid smaller than 0.46 pixels. In both cases, the aberrations are very small and have no influence on the determination of the centroids of the images. Finally, we analyzed the influence of the scattering at the surface have onto the evaluation. For this experiment, we observe that this effect can be considered as a very random error on the final evaluation of the surface.