WORLD WIDE WEB‐BASED INFORMATION STORAGE AND RETRIEVAL

Abstract

This paper describes the design and implementation of a system for computer generation of linked HTML documents to support information retrieval and hypertext applications on the World Wide Web. The approach is based on work by Salton and others, but extends the concept to be compatible with the World Wide Web browser environment by adding an interactive indexing technique that is well suited to the mouse‐based point‐and‐shoot input common to windowed browsers. The system does not require text query input, nor any client or host processing other than hypertext linkage. The goal of this work is to construct a fully automatic system in which original text documents are read and processed by a computer program that generates HTML files, which can be used immediately by Web browsers to search and retrieve the original documents. Thus, a user with a large collection of information — for instance, newspaper articles — can feed these documents to the program described here and produce directly, without further human intervention, the necessary files to establish World Wide Web home and related pages, to support interactive retrieval and distribution of the original documents.