Clarify svd documentation

[eschnett]

The svd(A, B) documentation (and similarly the svdfact documentation) currently say

  svd(A, B) -> U, V, Q, D1, D2, R0

  Wrapper around svdfact extracting all parts the factorization to a tuple.
  Direct use of svdfact is therefore generally more efficient. The function
  returns the generalized SVD of A and B, returning U, V, Q, D1, D2, and R0
  such that A = U*D1*R0*Q' and B = V*D2*R0*Q'.

I'm familiar with SVDs in the mathematical sense, but not with Julia's implementation. From convention, I know what U and V are, but I don't know the other symbols (Q, D1, D2, and R0). I assume that at least one of these is a vector (i.e. a diagonal matrix). I also assume that some additional conditions hold between these quantities.

If you're familiar with SVDs, could you update the documentation to have more detail? For example, a description in the style of svd(A) would already be sufficient for me.

[jiahao]

The notation is for the generalized SVD as provided by LAPACK.