Structural Types

Finally, before we take a look at the binding rules that tie this type system together, we will explain one last feature that takes algebraic types to the next level: structural types. These allow you to create aliases for types, but these are allowed to be recursive so you can create data structures from them.

Syntax

To talk about structural types we need to take a look at the syntax Ryna uses to define them:

type Name = AnotherType;

In this example, Name would be the generated structural type and AnotherType would be the so called alias, which is the "translation" of the type Name. Of course, these could also be made using generics:

type Name<T1, T2> = AnotherType<'T1, 'T2>;

Recursive Data Types

This is fine for having shorter names for complex types, but not much else. The power in these comes from allowing recursion. This would be an example of a recursive structure that can encode a binary tree of Int:

type BinaryTree = () | (Int, BinaryTree, BinaryTree);

This is a common introduction to these data types. In this case, a binary tree is either an empty tree (represented by a ()) or an Int followed by two subtrees with the left and right branches. Of course, this is not the only way to construct such an structure, but it is a valid example.

Now that we have seen every feature of the type system we can define the binding rules.