Imprecise documentation for solving linear equations.

Hi, thanks for the crate. I'm new to Rust, so this request might not have much sense. I was looking for some linear solvers but encountered the following problem.

In the description of the functions `.solve()` and `.solve_mut()` on the [website](https://nalgebra.org/docs/user_guide/decompositions_and_lapack/) omitted conditions on the matrix `A` which has firstly to be decomposed. “where `A` is represented by `self` and `x` the unknown”, `self` is more likely to be understood as a matrix itself rather than its factorization. If one tried to read the [documetation](https://docs.rs/nalgebra/0.29.0/nalgebra/base/struct.Matrix.html) for details, then it is not clear where to find a `.solve()` method. After implementing the method to the matrix directly, the only received message is “no method is found for this struct”.

# Request

It would be helpful to see:
1) a mention on the [userguide website](https://nalgebra.org/docs/user_guide/decompositions_and_lapack/) of the decomposition as a requirement.
2) Alternatively (or in addition), a default decomposition could be implemented for a `matrix.solve(&vector)` with a warning from the compiler that the decomposition is default or is required.
3) Also, to add a mention of these methods on the [website](https://docs.rs/nalgebra/0.29.0/nalgebra/base/struct.Matrix.html) since it is not clear where to look for them.

# Details
Here are the default choices for the decomposition in other languages:
- Julia
- [LU default](https://docs.julialang.org/en/v1/stdlib/LinearAlgebra/#Base.:\\)
- Python numpy
- [No idea, reference to LAPACK](https://numpy.org/doc/stable/reference/generated/numpy.linalg.solve.html?highlight=solve#numpy.linalg.solve)
- R
- [No idea, reference to Fortran](https://www.rdocumentation.org/packages/base/versions/3.6.2/topics/solve)
- ndarray
- [only LU (and by default)](https://docs.rs/ndarray-linalg/0.14.1/ndarray_linalg/solve/index.html)

# Help
I would be happy to help making any changes to make the description more understandable.