In this paper, we present an approach for combining different document representations to support retrieval systems to deliver similar documents from different views.