Cleaner matrix arithmetic

[jdwhite48]

Since Matrix<_> (i.e. Vec<Vec<_>>) are variable-length, this leads to a number of issues wherein I must implement my own custom functions for matrix arithmetic, instead of being able to use the syntactic sugar provided by Add, Mul, etc. that would make the matrix representation easier to work with.

• [x] Figure out how to statically encode Matrix such that one can meaningfully define One (the identity matrix, if it's square; if not, ...?) and Zero (zero matrix).
• [ ] Restructure Matrix trait to use Add<Self>, Neg<Self>, Mul<Other>, etc.

[jdwhite48]

I gave up on implementing a custom Matrix type myself and will be pulling from the nalgebra library to whatever extent possible. See 12-integrate-matrix-libs/src/data_structures.rs and #12 for more details about how this affects encoded Matrices (TL;DR fields are amenable to nalgebra, but group action / scalar multiplication still requires a custom implementation).