Inherited from #17 Current code only supports real numbers, i.e. `f32` and `f64`. Some algorithm in the linear algebra are closed in real number, but others are not.
