Matthew Kay

> That is, there is a danger of endless work to extend support for weighted draws if one part works too well. I still think support for weighted draws would...

That's an interesting solution @n-kall. @paul-buerkner would a default approach for weighted draws be to pass through a weights argument? Do most of the diagnostic functions have `...` in them...

> Review of parse_variable_indices.R and, if it looks good, rewriting as_draws_rvars.draws_matrix() to rely on parse_variable_indices Thanks @jsocolar! I added (hopefully not too many :) ) comments with some suggestions for...

Wonderful to hear! I think that the variable index parsing aspects of this PR may also end up being related to some of the conversation in #187 about exposing internal...

I'd be curious if anyone out there knows anything about the literature on HDI estimation methods? In {ggdist} I've been using the two methods from {HDInterval} and the more I...

Maybe this is the wrong place for this, but I wonder if part of putting those generics here might include some more standardization in their interfaces. At present there are...

Following on this issue from brms (https://github.com/paul-buerkner/brms/issues/552), I think it would be nice to make sure these interfaces have enough standardization that we could add a marginalization interface within posterior...

This is a really tough call! I think my instinct is to have fewer redundant ways of doing things if at all possible (lots of friction happens from a user's...

Ah yeah, that makes sense. Some of those functions are already covered, including everything in the [Math generic](https://stat.ethz.ch/R-manual/R-devel/library/base/html/groupGeneric.html), which includes `sin()` --- but I see that your point is not...

Yeah, thanks that would be helpful! I think it's mentioned in the vignette but probably not elsewhere. Alternatively, we could discuss if we want to keep that difference? It may...