Toward a 5M model of digital libraries

Whereas the DELOS DRM and the 5S model of digital libraries(DL) addresses the formal side of DL, we argue that a parallel 5M model is emerging as best practice worldwide, integrating multicultural, multilingual, multimodal digital objects with multivariate statistics-based document indexing, categorization and retrieval methods. The fifth M stands for the modeling the information searching behavior of users, and of collection development. We show how an extension of the 5S model to Hilbert space (a) points toward the integration of several Ms; (b) makes the tracking of evolving semantic content feasible, and (c) leads to a field interpretation of word and sentence semantics underlying language change. First experimental results from the Strathprints e-repository verify the mathematical foundations of the 5M model.