Carlos Parada

What would such an interface entail? I ask because, apart from MCMCChains, I don't think there's any other that you'd need to using an interface, and it's already easy enough...

> > What would such an interface entail? > > I ask because, apart from MCMCChains, I don't think there's any other that you'd need to using an interface, and...

> But even ArviZ falls short in the storage area, because it's still biased toward arrays. But in principle, one can have parameters be any Julia object (e.g. Distributions.LKJCholesky produces...

> So something like StructArrays (or ComponentArrays) is one way to go. StructArrays doesn't do recursive conversion though, so you'd need to tell it to on a per-type basis, which...

(I'm unable to figure out where "DataTypesBasic" is actually being used; if it's not being used, it can be dropped to reduce compilation time.)

The compile time issue seems to have gotten significantly worse, as it now takes about a second to load: ``` julia> @time using WhereTraits 0.948938 seconds (3.40 M allocations: 239.641...

I'm not sure how much work has been done on this already, but I think we might want to consider whether we want to use Plots.jl or a package that...

> I appreciate the pointer to DensityInterface, but it doesn't look like that package has a `logpdf` function. I'd be happy to adopt a common interface, but I don't think...

Hopefully someday my brain will stop making off-by-one errors, but today is not that day.

> Maybe I need to rephrase. After your change, how do you determine the largest value that is smaller than the tail values? If you don't sort them, I think...