A stepwise distribution, e.g. in bins of luminosity and redshift, can be implemented in the jollof methodology. This could dramatically increase the number of parameters, since there will be one parameter for each bin. That said, this kind of thing has been done often, e.g. for estimation of expoplanet distributions.
Another question is whether the bin edges should be fittable as well, but for a first version they could be fixed.
A stepwise distribution, e.g. in bins of luminosity and redshift, can be implemented in the jollof methodology. This could dramatically increase the number of parameters, since there will be one parameter for each bin. That said, this kind of thing has been done often, e.g. for estimation of expoplanet distributions.
Another question is whether the bin edges should be fittable as well, but for a first version they could be fixed.