Many areas in computational science and engineering require geometric software. Areas like solid modeling, CAD/CAM, computer graphics, image processing, computational cartography and robotics deal with geometric domains and invariably need good quality software. For example, mesh generation in finite element methods needs geometric software to decompose a geometric domain into simpler elements such as tetrahedra and hexahedra.
Although not as per demand, quite a few geometric software packages exist commercially. Unfortunately these packages suffer from two major drawbacks: inefficiency and unreliability. These are an immediate fall-out of poor-quality algorithms that employ ad hoc methods to reach the solution, which are otherwise avoidable by careful research. For example, commercially available software packages for mesh generation often cannot handle complicated geometric domains and occasionally produce poor quality meshes. On top of this there is the added difficulty of maintaining robustness of these packages under numerical errors caused by finite precision arithmetic. This problem is really serious for geometric software. Since standard geometric algorithms rely on the fact that the combinatorial component is consistent with the numerical component, they fail when this premise becomes false due to numerical errors. Almost none of the commercially available packages deal with this problem.
In the area of Computational Geometry, research in geometric algorithms has progressed with tremendous pace in the 1980's and 1990's. Unfortunately, little collaboration took place between the industries dealing with geometric software and the research community in this field. As a result, little percolated from research articles to actual software. Very recently a new thrust has been adopted by the computational geometers to convert their algorithms into real quality software. Several universities and research centers abroad have taken up projects in collaboration to produce such geometric software. For example, the CGAL project --- a collaboration between Max Plank Institute, Germany, Utrecht University, Netherlands and Tel Aviv University, Israel --- aims to produce a library of geometric tools which can be used and developed further by the industry.
Quite a few geometers are scattered all over India in different universities and research organisations. The main purpose of this article is to appeal for a collaborative effort to produce quality software so that national industries can use them as a platform for further applications. The author of this article believes that such an effort is necessary. However, it is not yet clear how we should proceed to make such an effort successful. Given difficulties in communication due to poor Internet connectivity, this collaboration is doomed to face many problems. On the other hand, the time is ripe for such adventures and we should not wait further. Also, overall communication is improving. A serious discussion should be started over the net on this issue. This author encourages concerned people to send their views on the subject either by postal mail or electronic mail.