Abstract. Ontology languages are being proposed to provide machine-understandable descriptions of resources that permit easy location of these resource. Content managers can also benefit for similar techniques as they also have to cope with large number of resources, in this case, documents. Moreover, the new crop of document processors (editors, workflow, content managers) are metadataaware. Each processor takes the document as input, processes it, and potentially, attaches, either manually or automatically, some metadata. For large and heterogeneous document sets where metadata come from diverse origins, automatic assistance is required to map and keep consistent the ontology layer with the resource layer. This paper describes how the effects of the distinct processes (e.g. editing, retrieving, consulting) on the document repository can be automatically propagated to the ontology model counterpart. To this end, a rule-based approach has been used. Their self-contained, isolated and...