Refactoring of analysis code
Hi @npostiau, @hanwen, @lyuan,
As I mentioned at our last meeting, I think the analysis code would profit a lot from a refactoring. Its main goals would be to separate the physics analysis from the event looping, switch to a class-based description for main physics objects, and clean up the top-level physics analysis code by moving all the technical and low-level details to dedicated classes. I prepared a wiki page with the description of the proposal. In fact, it includes two somewhat related approaches, each of which is based on a real-life analysis. I tried to give some examples so that the discussion is not too abstract.
I would like to ask you for a feedback about the proposal, including which of the two approaches you prefer and if you have something better in mind. If you see cases where the proposed approaches would not be appropriate, I would also appreciate this.