The typical meaning of M^(-2) would be inv(M)^2. It would seem reasonable to extend pow(M,n) to negative n in this way. There doesn't seem to be any risk; pow(M,-2) might throw an error if M is noninvertible, but it always throws an error at the moment. Unless some problem with this extension is foreseen, I will extend the definition of pow(M, n) in this way when I get a chance.
The typical meaning of M^(-2) would be inv(M)^2. It would seem reasonable to extend pow(M,n) to negative n in this way. There doesn't seem to be any risk; pow(M,-2) might throw an error if M is noninvertible, but it always throws an error at the moment. Unless some problem with this extension is foreseen, I will extend the definition of pow(M, n) in this way when I get a chance.