Binned vs Unbinned dist

[jonas-eschle]

Should we have a distinction between the two? And how do they react on data?

Unbinned with unbinned is clear

Binned acts more as a "parametrized histogram": unbinned data is evaluated in a step-func like manner (makes sense), but what should a hist into a binned one return? Histogram-like (as in zfit1?)
But what is binned in unbinned? I.e. in a poisson? Not possible, because not clear? (some may want the poisson of counts, others of bincenters? Also, needs to broadcast with params? -> let user decide and feed in the flattened counts?

Need for "diag PDF", i.e. Diag(Poisson)?

What should "space" be called, "domain"?