Words of wisdom or famous last words? I have used IDL's curvefit.pro in the past, but have run into problems when I want to fix the value of a certain variable--or constrain the value of a variable. Also, the uncertainties of the fitted parameters quoted by curvefit are often very misleading (you may argue I don't know what I'm doing). I have found that assigning point weights seems to sometimes cause problems for curvefit.pro.

Marquardt wrote mpfit.pro, which resolves all these issues. I thought you might need it one day, so here's the programs and a nice tutorial: http://cow.physics.wisc.edu/~craigm/idl/idl.html

No comments.