A sum()method is already planned for the next release (which will happen on Febuary 3 2019). In the mean time you can do: vector.iter().fold(0.0, |a, b| a + *b).
Multiplication by a reference to a scalar is not implemented because it didn’t seem useful since the scalar types are expected to be small in size and must be Copy anyway. Is there a specific use-case where multiplying by a ref is much more convenient than just deferencing the scalar if it’s a ref?