Daniel Schwabeneder

Yes, but I think then we have to change the type recipe for distributions. I have not thought about it in detail, though.

Nice.

That could/should be done for contour, surface and wireframe as well.

Can't it be a standalone package?

I like it. So StatsPlots would only keep the recipes that depend on StatsBase, Distributions, etc.?

Could you maybe add a note about this in the boxplot section of the README?

OK sure, then let's wait a little longer for other comments and opinions.

Does it make sense to use `barwidth` for this?

I don't think they can be controlled seperately within one series. A bigger barwidth implies smaller spacing and vice versa, no?

> After updating Plots v1.0.7 to v1.0.8, I now get the following error: Thanks for reporting, that should be fixed in https://github.com/JuliaPlots/RecipesPipeline.jl/pull/43,