Introductory documentation

Start by reviewing the existing reference documentation, the concerns in issues #160 and #327, and the ndarray for NumPy users guide cited in the issue. Done would be a separate high-level overview that explains ndarray's structure and design principles while gradually introducing the library to newcomers.

I was reasoning around this during the last few days while working on my PRs (and reading some issues like #160): it is really complicated to get a good overview of the way ndarray is structured solely relying on the reference documentation. It is even more complicated when the method you are looking at calls functions that are annotated with #[doc(hidden)] (it happens quite often in the core structs and traits implementations)

Without @jturner314's ndarray for NumPy users it would have taken me at least twice as long to reach my current level of understanding (which is still very partial). This echoes #327, for instance.

Would it be of interest to produce some kind of separate documentation, similar in spirit to http://nalgebra.org/ for nalgebra or http://scikit-learn.org/stable/ for scikit-learn?
A high-level overview of the overall crate, aimed at explaining the underlying design principles, gradually introducing a newcomer to the library.
I would be happy to work on it, at least to further my own understanding of ndarray.
What are your thoughts on the matter @bluss?