Hello! I've been working on a lot of projects having to do with general algebra in Swift dealing with groups, rings, fields, etc, as well as a library for fast matrix computations that allows for parametrization of any type whose values can be thought of as elements of a ring. I was wondering if the Swift Numerics library would have any use for this? Currently I've been trying to just work on it to implement different lattice cryptography schemes, but as I've been working on it, I've found myself adding more functionality to my BigNumber and MatrixKit libraries, and it's developing into more of a general algebra library. I've been working on making it so that the matrix operations are automatically optimized depending on the parametrized type.
For example, Matrix<Double> uses Accelerate to compute its computations, and I'm working on other optimizations for Matrix<Int> and Matrix<BigNumber>, etcetera.
Is this the kind of thing that would be useful to the Swift Numerics library? Or is this better fit elsewhere?
Future expansion may assume the availability of other standard interfaces, such as BLAS (Basic Linear Algebra Subprograms) and LAPACK (Linear Algebra Package), but modules with more specialized dependencies (or dependencies that are not available on all platforms supported by Swift) belong in a separate package.
AFAIK Accelerate is not available on Windows nor Linux.