1. 17

An extremely small and incomplete linear algebra I wrote over the weekend to test Rust’s handling of const generics.

  1.  

  2. 3

    I’m so excited about const genetics coming together finally. Lack of them pushed me away from Rust like 18 months ago.

    1. 1

      Yes, they are little unstable but on the whole very functional. I’m excited too.

    2. 2

      How does/will it compare with existing rust linear algebra libraries such as rulinalg and nalgebra?

      1. 5

        compared to rulinalg, aljabar is purely array based, and therefore all values will be stored on the stack. Also, in aljabar the size of the vector/matrix is fixed and the compiler will prevent what would otherwise be runtime errors, such as multiplying two matrices of incorrect width/heights. aljabar is pretty similar to nalgebra, except nalgebra requires a specific special type to be specified to each number that the width and height could be. This is a much simpler interface that wasn’t possible until recently.

        1. 1

          Nice. I tried both of them and wasn’t really satisfied by their respective interfaces. I hope that you’ll succeed at creating a more straightforward library :)