dredging up the past

I'm so chuffed! Someone actually found something useful in my thesis. This is surprising because the experiment that it's based on has finished, and the chap who contacted me works on my new experiment. However some of the statistical techniques (which my thesis is chock full of) are obviously still relevant. I guess he found it with a Google search.

Dear Tim,

i'm puzzled about a minimization problem performing a log-likelihood fit
with MINUIT (Dalitz plot fitting) that you seem to have solved in the past
while working at your PHD thesis (pag.121-122).

In these pages you say that under a certain cut-off you used a polynomial
function instead of logarithm and that you matched value and first two
derivatives at cut-off. The idea of this strategy looks to me very smart
but i was wondering how did you performed this matching "talking" to
MINUIT....

Any suggestion or help is greatly appreciated.
Thank you in advance.
Cheers,
Alexis

I dug out the piece of code from my old analysis and sent it back to him with some explanation. He'll have to convert it from Fortran to C++, but it's mainly just the concept and a bit of maths, so that's no big deal.