Giulio Valentino Dalla Riva

Given the lack of response, I would say so. It would be great if some of the developers took the time to archive the repo and pull down the website.

> > if you can put in a PR together that is non-breaking with respect to current api that's totally fine. If its breaking just increment version in Project.toml >...

The need for types (sorry) comes from trying to have multiple seasonality effects. Each seasonal component may have its own _ar_ and _ma_ coefficients. One $(2,0,3)s$ component, for example, will...

Absolutely right. I'm using static arrays to store the coefficients as well ☺️ (with a convenience constructor function that takes a regular vector and convert it to a SVector before...

Something similar seems doable. With the minor cavaet that, for the moment, in `Stranbo` the functions `sarima` / `sarimax` construct the components, they don't run the simulation (that is left...

> What's the advantage of not running the simulation and making the user go the extra step of calling sample? Doesn't this unecessarily complicate things? I'm no sure. With an...

> Can we define the convenience sample(::Component) dispatch as well? Sure, that's already there: ```Julia julia> using Stranbo julia> test = sarima(ar = [.8], ma = [2.]) SARIMA{Float64}(1, 0, [0.8],...

Sorry, had very busy days, I'll get to this asap :-)

Hi Albert, many thanks for the answer at #164 :-) Would something like ```julia for i in eachindex(A) A[i] ~ Gamma(2.0,3.0) end ``` work, instead of broadcasting, in the example...

## Review checklist for @gvdr ### Conflict of interest - [x] As the reviewer I confirm that I have read the [JOSE conflict of interest policy](https://github.com/openjournals/joss/blob/jose/COI.md) and that there are...