This is just a placeholder to allow links from other issues and PRs.
Very generically, it is about "Event by event reweighting using matrix elements computed for different parameter values". Personally this is something I am very keen about for physics/statistics studies.
@oliviermattelaer, anyway I know this is not a priority and we should discuss this more in detail eventually, also because I am not 100% sure what MG5aMC allows now. This is something we vaguely discussed also with Stefan and Robert a couple of months ago.
Anyway, technically my point is that it would be nice to keep the handling of parameters flexible enough so that one can potentially have more than one "parameter set" inside a calculation of MEs. Maybe this is not needed and everything could/should be done by doing several runs with different parameter sets and writing results to file, but I like to think that we can have several parameters within the same c++ run. For instance, this would allow computing derivatives of matrix elements with respect to parameters...
Low priority, just a placeholder.