The Web, the largest unstructured database of the world, has greatly improved access to documents. However, documents on the Web are largely disorganized. Due to the distributed nature of the World Wide Web it is difficult to use it as a tool for information and knowledge management. Therefore, users doing the difficult task of exploring the Web have to be supported by intelligent means.This paper proposes an approach for document discovery building on a comprehensive framework for ontology-focused crawling of Web documents. Our framework includes means for using a complex ontology and associated instance elements. It defines several relevance computation strategies and provides an empirical evaluation which has shown promising results.