Intelink Automatic Link Generation

Although many tools have been developed to aid the conversion of pre-existing documents from proprietary work processor formats to HTML for access via the World Wide Web and related Intranet documentation access systems, such as the Intelligence Community's Intelink system, few tools exist to aid the automatic discovery and implementation of hypertext links between related documents. This proposal not only describes a technique to aid the document preparer in finding and adding links to related documents, but also to alert the document preparer to new links as additional documents are added to the document server. The described method uses a client-server architecture, where a server maintains a database of document descriptions, and a client formulates descriptions for new documents. The client's descriptions are submitted to the server, which returns pointers to documents that are likely candidates for linking to. The client then compares candidates against the primary document to further judge the merit of each candidate, and to determine how best to link the primary and candidate documents. The server uses full word indexing and a binary document classifier to aid in selecting target documents. The client uses a heuristic to choose hyperlink points.