Hi, see the issue on Kronecker.jl:
MichielStock/Kronecker.jl#137
This person wants to use a Kronecker matrix for a covariance matrix in a multivariate normal distribution (Distributions.jl). Since these use PDM, I think the best option would be to include a method for a Kronecker PDM here? Almost all functionality should inherit.
My other option would be that PDM(K) with K a Kronecker matrix yields a new type KroneckerPDM (Kronecker product of PDM) with the methods of this package copied here.
What do you think would be the best option?