Carlos Parada

Should we memoize `inv(σ)` since there's a good chance it'll be frequently used?

> > Should we memoize `inv(σ)` since there's a good chance it'll be frequently used? > > No, I don't think it's worth it. In my experience it just causes...

@devmotion I think the PR is mostly done apart from -- 1. Tests for the multivariate case (which I'm writing) 2. Whatever is wrong with the discrete CDF functions, which...

All tests now pass, except for `support` when called on normal distributions that have `ForwardDiff.Dual` numbers as arguments. I'm going to be completely honest and say I'm not sure what...

> I hope you don't take this personal but I don't think we can merge the PR in its current state. In my opinion it is difficult to review and,...

I also would like to point out there have been essentially no changes to the behavior, of `LocationScale`, with the exception of: 1. Bug fixes for the `location` and `scale`...

Ok; if you'd like to implement renaming and all the different changes in a different PR, go ahead, then. I'll make a single PR that only changes LocationScale to allow...

> I guess another important difference is that BAT explicitly distinguishes between prior and likelihood, and can support likelihoods that are not (explicitly) tied to data. The separation between prior...

> An alternative path would be to develop a separate package named SpecialFuns.jl (or any similar name), which supersedes Bessels.jl and related functionality without dependencies or any promise of backwards...

I'd love to see these; I think the most AoG way to write these would be by just adding to a 2d plot, e.g. ``` data(df) * mapping(:x, :y) *...