ConstTypeArgs.Numerics Library

[zacharylayne]

A ConstTypeArgs.Numerics library would incorporate some of the functionality found in the System.Numerics namespace. However, decisions need to be made over what functionality to bring over.

• Should the library provide const type arguments for one or more of these types? BigInteger, Complex, Matrix3x2, Matrix4x4, Plane, Quaternion, Vector<T>, Vector2, Vector3, Vector4.
• Are there const type interfaces needed? If so, what System.Numerics interfaces should be covered?
• What operations & methods should be incorporated?
  • Should operations & methods be part of one or more static utility classes, encapsulated in K_Delegate types, and/or static methods on const type arguments or argument holders?
  • When using const type arguments in operations & methods, is it possible or desirable to cache results?
  • Is it possible or desirable to calculate results during type initialization?
• Should there be new argument holders such as Positive<K>, Pow2<K>, and so on?
• Should there be a Number<K>, Integer<K>, FloatingPoint<K> and other argument holders to unify numeric types, such as byte, sbyte, int, uint, Int128, UInt128, long, ulong, short, ushort, decimal, double, float, and Half?
• Is there any utility in providing a discard argument (__) that unifies numeric discards found in other libraries? If so, should it be an abstract class or another interface?