# Convert between View Matrix and Tuple of Camera Position, LookAt Vector, Up-Vector

given a View-Matrix $M$ that can transform world coordinates into camera space, how can I convert between this representation and a more human readable form of Position ($\vec p$), Look-at vector ($\vec l$) and Up-vector ($\vec u$)?

Also are both forms interchangable, so can I represent every View Matrix as a tuple of $(\vec p, \vec l, \vec u)$? If not, what if the Matrix was created using only a series of translations and rotations?

Thanks!

-