Scientific computing over the Web can be applied to solve problems of many types, namely as appear in the activity of companies. In this study, the computation is done with two remote servers, without the user's perception, thus disclosing no clues on the computation. As an illustration, we make use of random sampling, which is needed in many activities, such as manufacture and research. Sampling is presented in its two basic versions, with and without replacement, the latter the typical choice in practice. Our objectives are: to emphasize the pertinence of the Web for scientific computing; to illustrate this computing with one of its countless applications; to show the possibilities of using remote, distributed servers; and to imply that this can provide the desired link between industry and the university.In our opinion, the full benefits of scientific computing over the Web have not been reaped, despite the availability of this communication medium.