[SciPy-user] Juicy tidbit from the ODRPACK User's Guide

[zunzun at zunzun.com]

I've been looking for a simple way to calculate a
covariance matrix for nonlinear curve fitting so
that I could add fit statistics to my web site,
and have been wargling back and forth over various
packages, files, languages, more files, etc. for
about two months now.  I *almost* started to add
additional ODR binding code to Robert Kern's
odrpack.c, when lo and behold;


On page 33 of the ODRPACK User's Reference Guide one finds:

 If MAXIT = 0 then no iterations will be taken, but whatever
 computations are required to complete the final computation
 report will be made. For example, by setting MAXIT = 0 and
 the third digit of JOB to zero, the user can compute the
 covariance matrix Vbeta for the input values beta and delta.


Ohhhhh, I get it - READ THE DOCUMENTATION!  Boy, do I ever
feel like a total dork.  On the other hand I've learned a
lot about interfacing FORTRAN and Python, many thanks to
Pearu Peterson and Robert Kern.

     James Phillips
     http://zunzun.com