🚀 go-pugleaf

As requested here is a brief summary of the various replies
I received over the last week.

     ----------------

1) Suggested algorithms:
    -Simplex and Marquardt-Levenberg from Numerical Recipes, adpated to
     the problem at hand
    -Using a maximum liklihood method
    -A few people mailed me lists of source code avaliable from 
     various archives, NetLib, and where-ever.  Since many of these
     have been posted, or pointers to them posted, I won't repost them.

2) Several people said that this sort of fitting is sensitive to 
   the initial guesses at the parameters so their initial values
   and search regions should close to the "true" values.

3) Don't try to fit too many Gaussians to the data.

4) Subjective "chi-by-eye" type fitting is subjective.  Just because
   one fit "looks" better than another doesn't mean that it actually is
   better.  

     ------------------

Again, thanks to those who responded.

steve