Thanks, I'll check lmdif.
As far as implementing the interface for reading the covariance matrix / jacobian, it would be a very welcome addition. Afterall, fitting parameters of a model function is a common task in experimental data analysis, and it is instrumental to know the fitted errors and the correlation between the fitted parameters.
regards, Stan
Stanislaw Szpala, Ph.D.
Robarts Research Institute
100 Perth Drive
London Ontario N6A 5K8, Canada
tel (519)663-5777x34213