we broke them into distinct items in the schema to not have mixed data+computation, which can be confusing