Relationships are keys to the semantics and hence the Semantic Web. One interesting way to exploit relationships is to link documents such that terms in the documents are semantically related through well defined relationships. In his 1945 article, Dr. Vannevar Bush posited the idea of creating a device capable of recording all of human knowledge. Dr. Bush suggested thatsuch a device would allow its user to traverse a space of documents by following a "trail of associations" in the user's mind. The assumption however is that links of the Web are meaningful to the user. This assumption certainly does not hold for the eight billion pages on the Web. We consider a scenario where the user has some prior knowledge of the domain. Ourapproach to building such a system relies on two aspects of Semantic Web: (a) semantic metadata for documents, and (b) populated ontology with relationship instances.