Scope of the Project

• The graded triangulation routine implemented here uses two main data structures for storage. The TRI structure [Nan95] is used for the initial Delaunay triangulation and the EDGE structure [Sha96] is used for the AFM. This leads to many redundancies in storage and inefficiency in coding of several routines due to constant interaction between the two data structures. A new data structure optimized for the above work will help greatly in speeding up the routine and reducing memory requirements.
• The efficiency of the routine for determining the background triangle in which the trial point is located can be greatly enhanced by using a better (although more complex) search algorithm. The present implementation is a brute force method which searches through the entire list of triangles for the location of the given point. The information of the neighbours of each triangle in the list stored by the TRI structure can be possibly utilized to our advantage here.
• The input routine can be extended to include B-spline surface (BSS) and other parametrically defined surfaces as well, as well as for closed surfaces.
• The algorithm can be extended to use the Jacobian (curvature) of the surface as an additional parameter for point insertion. This should give more control on the triangulation obtained, and hence a better mesh.