Separate tokenization and parsed structures

[200sc]

Right now every struct in the project has an underlying object with elements and properties, which are all results of tokenization-- the elements from these types should be parsed into typed structures on each type resulting from parsing.