Symmetric

[cc7768]

Hi. First of all, this is a great idea and an important package to have for efficiency purposes.

Just a quick question/formality --I think it should also be specified that these matrices are hermitian(symmetric) in order for the Cholesky decomposition to be meaningful (see wiki). I could just be unfamiliar with the uses of the Cholesky decomposition for positive definite nonsymmetric matrices, but wanted to mention it just in case.

Also, the majority of your cases seem like they will be covariance matrices which are symmetric so this isn't relevant in those cases, but could possibly matter in cases like this where I don't think that scale matrices for an Inverse Wishart need be symmetric.