[Help-gsl] fitting with poison statistics

[H . Scheit]

  Will there be support for fitting data with poison statistics in the
  gsl?  I.e.  no least-squares (gaussian statistics), but rather
  max-log-likelihood, with the likelihood function corresponding to a
  poison distribution.  Best with an option to iteratively update the
  uncertainties to be sigma_i^2 = y(x_i), with y() being the fitted
  function.  This would be very useful for fitting of low statistics
  data.

  Thanks, Heiko