Merit Meanderings  - December 12,2001:

• What is Merit?

  • Takes as input the ntuple of reconstructed events (ie output of pdrApp)

  • Applies a series of cuts designed to weed out the bad stuff and improve resolution

    • Level 1-3 trigger selection cuts

    • Various background cuts

    • Cuts designed to improve resolution

    • etc.

  • Bins the results according to section (Front/Back), energy and incoming angle

  • Outputs the "final" answer - resolution and effective area in bins of energy and angle for Front/Back

  • Example here
    

• What is the problem?

  • Merit is not really a analysis/development package...

  • Currently, cuts are developed outside of Merit and then plugged in once they are deemed to work.

  • Requires two versions of the same thing

  • Does not easily allow multiple developers...
    

• Why incorporate into root?

  • Allow cut developer to "see" what they are doing
    

• What does "incorporate into root" mean?

  • Run Merit from the root command line

  • Execute Merit cuts for a particular event to determine if event passes or fails

    • Useful for applying current cuts when developing a new one

  • "Easily" produce plots in root display windows
    

• Where does it stand?

  • Merit integration into root not as easy as at first thought

    • Base objects must inherit from TObject

    • To make accessible to CINT, all higher objects need to get entered into the "dictionary" which was only successful after much hand editing of the results of dorootcint

    • root does not play well with the standard template library (at link time)

    • Recovery of objects from folders does not work quite the  way expected

    • Is it me, or do the root people reinvent everything? 

    • etc.

  • Current highly hacked version of merit does work. I am now trying to make "pretty" output plots for people to look at (see examples below).

  • If we really want something like this we should rewrite merit in the context of root
    

• Some examples produced from a macro written to run on the root command line

  • A basic set of PSF plots in bins of energy for the nearly downward going gammas
    
    

  • The summary of the PSF vs energy for the four angle bins:
    
    

  • Here is an attempt at a surface plot: