Cities are documented by large and heterogeneous multimedia corpora, including photographs, maps, plans, reports, videos, and web resources. While versioned urban models have been studied for representing the temporal evolution of city objects, the documents describing these objects are still often treated as static attachments. This paper investigates how versioned cities and versioned documents can be modelled together by treating urban documentation as principal evolving entities connected to evolving urban features. We propose a CityRDF extension as an OWL ontology, interoperable with CityGML 3.0 that represents documents, their qualified links to urban objects, and their associated temporal semantics within the same framework. The key idea is to separate the identity of a document from the interpretative relation that connects it to the city model, making it possible to distinguish document history, depiction time, and city-object history while preserving the version history of urban objects. The proposed model treats documents as core urban elements and links diverse, versioned records to equally versioned cities across time and scale for heritage analysis and urban exploration.