The problem of fitting a smoothing or interpolating curve to a discrete set of data values arises in almost every field of scientific research. An important class of curve fitting problems requires the determination of a smooth interpolating curve which preserves the shape of the data. An effective means of producing such a curve is provided by tension splines -- a generalization of cubic splines which includes additional parameters having the capability of controlling the behavior of the curve. The investigator has developed algorithms which relieve the user of the burden of having to experimentally determine appropriate values of these parameters. The algorithms automatically select the optimal parameter values required to satisfy user-specified constraints. A major deterrent to widespread use of this method is the difficulty associated with its implementation. The goal of the proposed research is a software package which implements the algorithms referred to above, thus providing a convenient, powerful, and versatile curve fitting method. The software is to be placed in the public domain and distributed to the scientific community through the ACM Algorithms Distribution Service.
