I so overwhelmingly agree. One thing I do dislike about distributions is the under-use of inner constructors.

Like why can’t the call be on a one dimensional vector instead of having the calculate mu and sigma and then put them in?

Not a big deal, but does present some interesting ideas.